• type 2 diabetes mellitus;
  • glipizide;
  • meta-analysis


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Declaration of interests
  9. References

This review aimed to address effects of sustained-release versus immediate-release glipizide on glucose control, insulin secretion, and compliance. We searched Medline, EMBASE, the Cochrane Library, and Chinese Biomedical database from inceptions to May 31, 2011, screened reference lists of relevant studies, and contacted pharmaceutical companies. Randomized trials and cohort studies were included. We pooled data using a random-effect model. Nineteen trials involving a total of 1440 patients and 2 retrospective cohort studies with a total of 13452 patients were included. Trials were of low quality. No trials reported patient important outcomes. The reduction of fasting plasma glucose from the baseline appeared larger for sustained-release than for immediate-release glipizide (mean difference −0.26 mmol/L, 95% CI −0.52 to −0.01). The reduction was not significantly different between the two drugs for HbA1c (−0.03%, −0.20% to 0.14%) or 2-hour postprandial plasma glucose (−0.21 mmol/L, −0.96 to 0.55). Sustained-release glipizide appeared to reduce insulin secretion from the baseline, whereas the immediate-release formulation increased the secretion (fasting insulin: −1.04 vs. 0.88 μIU/ml; 2-hour postprandial insulin: −2.94 vs. 0.24 μIU/ml). Patients administering sustained-release glipizide had less hypoglycemia (Peto odds ratio 0.21, 95% CI 0.08 to 0.52) and lower missed dosing (Peto odds ratio 11. 42, 95% CI 6.47 to 20.18). The cohort studies showed patient compliance results consistent with those of the trials. Sustained-release glipizide appears to achieve similar glucose control with decreased insulin secretion, fewer hypoglycemic episodes, and higher patient compliance than immediate-release glipizide. However, these findings are inconclusive due to inadequate study quality, short follow up, and unavailability of patient important outcomes.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Declaration of interests
  9. References

Type 2 diabetes mellitus is a serious public health problem, and is associated with a range of serious complications which result in reduced quality of life and premature mortality (1). An estimated 285 million people—about 6.4% of the world's population—had diabetes in 2010, and this number is expected to increase to 439 million (7.7%) by 2030 (2). Of all patients diagnosed with diabetes, 85% or more have type 2 diabetes (2).

Oral glucose-lowering drugs represent a milestone in the management of type 2 diabetes. These drugs are used when behavioral interventions alone fail to control glucose level. Second-generation sulfonylureas are recommended as a first-line therapy, particularly for those who are not overweight (3). Immediate-release glipizide (GIR) is an early second-generation sulphonylurea drug, and has been in clinical use for four decades (4). Sustained-release glipizide (GSR) was developed in the mid 1990s; its particular advantages lie in rigorous stabilization of drug plasma concentration and preferable sensitization of insulin, leading to optimal release of insulin (5–10). Another advantage of the sustained-release form is the simplicity of administration (i.e., once daily).

Randomized trials of glipizide versus placebo suggest that once-daily sustained-release glipizide effectively lowers glucose and improves insulin sensitivity with minimal adverse effects (5–8). However, studies comparing sustained-release versus immediate-release glipizide have been small, with discrepant findings (9–10).

Strict compliance with therapy is of paramount importance to glucose control and avoiding complications in type 2 diabetes mellitus. Non-compliance, however, occurs in 50–80% of patients (11). While many factors can influence drug compliance, simplicity of dosage regimen and adverse effects are two key therapy-related factors affecting compliance (12). So far, however, there are no compelling results demonstrating a compliance advantage for the sustained release form.

We conducted a systematic review and meta-analysis to address whether sustained-release glipizide results in better glucose control, fewer adverse effects, and higher patient compliance.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Declaration of interests
  9. References

Study Search and Selection

We searched Medline, EMBASE, the Cochrane Library, and Chinese Biomedical database (CBM) from their inceptions to May 31, 2011, and screened the reference lists of eligible trials and reviews. We contacted Pfizer for unpublished studies. We developed the search strategy in consultation with an information expert, and used the key words: “diabetes mellitus,”“diabet*,”“randomized controlled trials,”“random*,”“glipizide,”“compliance,”“adherence,” and “persistence.” No language or publication status restrictions were used.

To compare the effects of the two forms of glipizide, we included parallel, randomized controlled trials (RCTs) with a head-to-head comparison of sustained-release versus immediate-release glipizide in adult patients diagnosed with type 2 diabetes mellitus. We excluded crossover studies because of concerns about the carry-over effect. To assess compliance with drug administration, randomized trials and prospective and retrospective cohort studies were all eligible.

Pre-specified patient important clinical outcomes included diabetic neuropathy, retinopathy, nephropathy, erectile dysfunction, coronary heart disease, myocardial infarction, stroke, peripheral vascular diseases, cause-specific mortality, and total mortality. Pre-specified surrogate efficacy endpoints included HbA1c, fasting and 2-hour postprandial plasma glucose, fasting and 2-hour postprandial insulin, BMI, body weight, and lipid levels. Drug adverse effects included hypoglycemia, gastrointestinal adverse reactions, and any adverse event. Since there is no universally accepted definition of patient compliance, we used definitions reported by the trials themselves.

Study Screening, Assessment, and Data Extraction

We used standardized, piloted forms for study screening, risk of bias assessment, and data extraction. Two reviewers (QY and HL) screened the studies for eligibility and extracted data independently. Any discrepancies were resolved by a third party (LW).

We recorded study design, patient characteristics, interventions, and outcomes. For the surrogate efficacy outcomes (i.e., continuous outcomes), we extracted, for each group, data on the mean change from baseline to the final follow up and the standard deviation (SD) of the mean change; if this data was not available, we used the following formulae to calculate the mean and SD of the change:

Mean of the change = mean at the final follow up – mean at the baseline

  • image

SD1 is the SD at the final follow up, and SD2 is the SD at the baseline. r is the correlation coefficient, for which we used an estimated correlation coefficient of 0.4 (13)

We used the following items to assess the risk of bias for each randomized trial: allocation concealment, blinding (of patients, health care providers, data collectors, data analysts, and outcome assessors), completeness of follow up, and adherence to the intention-to-treat principle. We used the Newcastle-Ottawa Scale (NOS) to assess the quality of cohort studies, including selection and comparability of exposed and non-exposed cohorts, outcome assessment, and adjustment analysis on important baseline characteristics (14).

Data Synthesis

We employed the DerSimonian and Laird random-effect model to pool studies. We reported risk ratio and its 95% confidence interval (95% CI) for patient compliance data, and used the Peto odds ratio for adverse effect data, given the low frequency of events in trials (15). We reported the mean difference of the change and 95% CI for continuous data. For each trial, the change represents the difference of measures from the baseline to the final follow up.

We used the I-square statistic and heterogeneity chi-square test to examine statistical heterogeneity across trials. I-square represents the percentage of total variation across a trial that is due to heterogeneity. We used a small number of pre-specified hypotheses to explore heterogeneity, including administration methods (fixed-dose vs. titrated dose of two glipizides) and treatment duration (8 weeks vs. 12 weeks). An interaction test was used to compare the difference between two estimates of treatment effects in subgroups (16). We tested for publication bias using funnel plots and statistical tests (17).


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Declaration of interests
  9. References

Study descriptions

Twenty-one studies, 19 randomized trials (18–33) and 2 retrospective cohort studies, were included in our analysis (Figure 1). All 19 trials reported efficacy outcomes (Table 1). Eight trials (19, 21, 24, 26, 28) and the two cohort studies (37–38) addressed patient compliance. The diagnostic criteria for diabetes varied across studies. Twelve (23, 25, 27, 29–36) used the WHO 1999 criteria (39), one (21) used the American Diabetes Association 1997 criteria (40), and two (18, 22) used the WHO 1985 version (41). The remaining seven did not report their diagnostic criteria.


Figure 1. Flow chart of study selection

Download figure to PowerPoint

Table 1.  Characteristics of included studies
StudySample sizeMean age (years)Mean HbA1c at baseline (%)Mean FPG at baseline (mmol/L)Mean BMI at baseline (kg/m2)Diagnostic criteriaDuration of diabetes (years)InterventionsBaseline medicationOutcomes
  1. Notes: NR=not reported; ADA=American Diabetes Association; GSR=sustained-release glipizide; GIR= immediate-release glipizide; wks=weeks, mths=months; FPG=fasting plasma glucose; 2hPPG=2-hour postprandial plasma glucose; HbA1c= glycated hemoglobin A1c; FIns= fasting insulin level; 2hPIns=2-hour postprandial insulin level; AEs=adverse events; *= cohort studies; †= range

Zhang 2000[18] 6053.9 7.1 6.624.2WHO 19856.4GSR 5-20 mg/d vs. GIR 15-30 mg/d for 12 wksMetformin or acarboseFPG, 2hPPG, HbA1c, FIns, AEs, dosage
Wang 2001[19] 4057.28.2 8.8NRNR 6.6GSR 5-15 mg/d vs. GIR 15-30 mg/d for 12wksMetformin or acarboseFPG, HbA1c, FIns, lipid, AEs, dosage, compliance
Zeng 2001[20] 4253-74† 7.7 8.6NRNRNRGSR 5 mg/d vs. GIR 15 mg/d for 12 wksMetformin or acarboseFPG, 2hPPG, FIns, 2hPIns, HbA1c, C-peptide
Chen 2003[21] 4557.6 7.7 8.6NRADA 1997 3.66 GSR 5 mg/d vs. GIR 15 mg/d for 12 wksNRFPG, 2hPPG, HbA1c, FIns, 2hPIns, C-peptide, compliance
Guo 2004[22] 6457.6NR12.8 25.1WHO 19859.2GSR 10-25 mg/d vs. GIR 30 mg/d for 8 wksNRFPG, 2hPPG, FINS, TG, TC, AEs
Li 2004[23] 6551.6 7.3 8.322.9WHO 19994GSR 5-15 mg/d vs. GIR 15-30 mg/d for 12 wksMetforminFPG, FINS, lipid, HbA1c, BMI
Shen 2004[24] 6060 8.9 9.6NRNRNRGSR 5-15 mg/d vs. GIR 10-40 mg/d for 12 wksNRFPG, 2hPPG, HbA1c, AEs, dosage, compliance
Ji 2005[25]118 58.2 9.3 9.924.6WHO 19994.7GSR 5-20 mg/d vs. GIR 10-30 mg/d for 12 wksMetformin or acarboseFPG, 2hPPG, HbA1c, FIns, C-peptide, AEs, dosage
Wang 2005[26] 6060 8.9 9.6NRNRNRGSR 5-15 mg/d vs. GIR 10-40 mg/d for 12 wksNRFPG, 2hPPG, HbA1c, AEs, compliance
Yang 2005[27] 7044.8 9.2 9.5NRWHO 1999NRGSR 5-15 mg/d vs. GIR 10-30 mg/d for 8 wksNoFPG, 2hPPG, HbA1c, lipid, AEs
Hsieh 2006[28] 5751.5 7.71125.2NR6.25GSR 10 mg/d vs. GIR 10 mg/d for 12 wksNRFPG, HbA1c, AEs, body weight, compliance
Yao 2006[29] 9654.6NRNRNRWHO 1999 6.4GSR 5-15 mg/d vs. GIR 15-30 mg/d for 8 wksMetforminFPG, 2hPPG, HbA1c, FIns, C-peptide, lipid, body weight, AEs
Zhao 2006[30] 6056.1 8.8 9.626.9WHO 1999NRGSR 5 mg/d vs. GIR 15 mg/d for 12 wksNRFPG, 2hPPG, HbA1c, FIns, 2hPIns, C-peptide, AEs
Wu 2007[31]  78NR10.512.3NRWHO 199911-20GSR 5-15 mg/d vs. GIR 5-15 mg/d for 12 wksMetforminFPG, 2hPG, HbA1c, Fins,2hPIns
Zhang 2007[32]  71NR 8.1 8.99NRWHO 1999NRGSR 5-20 mg/d vs. GIR 15-30 mg/d for 12 wksMetformin or acarboseFPG, HbA1c, FIns
Cai 2008[33]  7643.69 8.3 8.324.6WHO 19993.02GSR 5 mg/d vs. GIR 15 mg/d for 8 wksNRHbA1c, FPG, 2hPPG, Fins, 2hPIns, lipid
Deng 2010[34]  5650.2 8.73 9.16WHO 19990.25-20GSR 5-10 mg/d vs. GIR 15-30 mg/d for 12 wksNoHbA1c, FPG, 2hPPG, adherence
Yao 2010[35]  6045.1 9.813.4  WHO 1999GSR 5-15 mg/d vs. GIR 15-30 mg/d for 12 wksNRFPG, 2hPPG, HbA1c, FIns, lipid, AEs, adherence
Zhang 2011[36] 262 55.1 WHO 19990.5-11GSR 5-20 mg/d vs. GIR 15-30 mg/d for 8 wksNRAE
Dezii 2002[37]* 992 18-64†NRNRNRNRNRGSR 15-30 mg/d vs. GIR 15-30 mg/d for 12 mthsNRPersistence and adherence
Harley 2002[38]*12198   52.1NRNRNRNRNRGSR vs. GIR for 54 mthsNoPersistence and adherence

Most trials failed to meet the criteria for methodological rigor (Table 2). In particular, very few trials reported whether they had performed blinding – and for which party – and concealed treatment allocation.

Table 2.  Methodological quality of included randomized trials
StudyAllocation concealmentBlindingLoss to follow upAdherence to ITT analysis
Zhang 2000[18]UnclearUnclear0Yes
Wang 2001[19]UnclearUnclear0Yes
Zeng 2001[20]UnclearUnclear0Yes
Chen 2003[21]UnclearUnclear/5Yes
Guo 2004[22]UnclearUnclear0Yes
Li 2004[23]UnclearUnclear0Yes
Shen 2004[24]UnclearUnclear0Yes
Ji 2005[25]UnclearUnclear0Yes
Wang 2005[26]UnclearUnclear0Yes
Yang 2005[27]UnclearUnclear0Yes
Hsieh 2006[28]Central randomizationPatients, healthcare providers1/7Yes
Yao 2006[29]UnclearUnclear0Yes
Zhao 2006[30]UnclearUnclear0Yes
Wu 2007[31]UnclearUnclear0Yes
Zhang 2007[32]UnclearUnclear/1No
Cai 2008[33]UnclearUnclear0Yes
Deng 2010[34]UnclearUnclear0Yes
Yao 2010[35]UnclearUnclear0Yes
Zhang 2011[36]UnclearUnclear0Yes

The two cohort studies selected their cohorts from managed care organizations in the United States. The outcomes were measured through the detailed claims and record linkage in the databases. The follow up periods were 1 and 4.5 years. Important baseline characteristics were comparable between groups, and analyses were conducted to adjust for the influence of these characteristics.

Comparative treatment effects

Of the 16 included trials, 5 used fixed-dose glipizide (GSR: 5 or 10 mg daily; GIR: 10 or 15 mg daily), while the other 14 used titration schedules (5–25 mg daily for GSR, 5–40 mg daily for GIR). Treatment duration ranged from 8 to 12 weeks. Table 3 presents the effects of GSR versus GIR. No trials reported long-term patient important outcomes.

Table 3.  Comparative efficacy of sustained-release versus immediate-release glipzide
 No. of studiesSample sizeMean change* from baselineMean difference GSR and GIR groups (95% CI)Heterogeneity
  1. Trials were pooled in random effect model.

  2. *change between end of follow up and baseline

Hemoglobin A1c (%)(18-21, 23-28, 30-35)161018−2.54−2.51−0.03 (−0.20 to 0.14) 0.650
Fasting plasma glucose (mmol/L)(18-28, 30-35)171082−2.09−1.83−0.26 (−0.52 to −0.01) 0.0835%
2-hour postprandial plasma glucose (mmol/L)(18, 20-22, 24-27, 30-31, 33-35)13 849−3.62−3.41−0.21 (−0.96 to 0.55)<0.00167%
Fasting insulin (μIU/ml)(18, 20-23, 29-34)11 713−1.04 0.88−1.91 (−3.59 to 0.24)<0.00183%
2-hour postprandial insulin (μIU/ml)(20, 21, 30-31, 33) 5 301−2.94 0.24−3.18 (−5.47 to −0.90) 0.1639%
Triglycerides (mmol/L)(22, 27, 29, 33-34) 5 362−0.48−0.13−0.35 (−0.58 to −0.12) 0.0267%
Total cholesterol (mmol/L)(22-23, 27,29, 33-34) 6 427−0.17−0.07−0.10 (−0.21 to 0.01) 0.970
Low density lipoprotein (mmol/L)(23, 27, 33) 3 211−0.09−0.11 0.02 (−0.05 to 0.10) 0.950
High density lipoprotein (mmol/L)(23, 27, 33) 3 211 0.01 0.07−0.06 (−0.12 to 0.00) 0.740
Glycemic control

Pooling of 16 trials of 1018 patients (18–21, 23–28, 30–35) addressing HbA1c suggested that the mean changes from baseline for GSR and GIR were not significantly different (−2.54% vs. −2.51%, difference in mean changes 0.03%, 95% confidence interval −0.20 to 0.14, Figure 2). Meta-analysis of 17 trials involving 1082 patients(18–28,30–35) showed that the reduction in fasting plasma glucose was slightly greater in GSR versus GIR (−2.09 mmol/L vs. −1.83 mmol/L, −0.26, −0.52 to −0.01). The reduction in 2-hour postprandial plasma glucose in 13 trials including 849 patients (18, 20–22, 24–27, 30–31, 33–35) was not significantly different between the two drugs (−3.62 mmol/L vs. −3.41 mmol/L, −0.21, −0.96 to 0.55)


Figure 2. Change in HbA1c from baseline between sustained-release and immediate-release glipizide

Download figure to PowerPoint

Insulin secretion

The pooled results of 5 trials of 301 patients showed that GSR reduced 2-hour postprandial insulin from the baseline by 2.94 μIU/ml, whereas GIR increased its secretion by 0.24 μIU/ml (mean difference −3.18 μIU/ml, 95% CI −5.47 to −0.90, Figure 3). Trials reporting fasting insulin suggested a similar trend in change from baseline (1.04 μIU/ml reduction in GSR vs. 0.88 μIU/ml increase in GIR, −1.91, −3.59 to 0.24). However, this result is apparently inconsistent across trials (heterogeneity p<0.001, I-square = 83%).


Figure 3. Change in 2-hour postprandial insulin from baseline between sustained-release and immediate-release glipizide

Download figure to PowerPoint

Blood lipid levels

Six trials of 427 patients (19, 22–23, 27, 29, 33–34) reported blood lipid levels. The change in lipids level from baseline was not significantly different between GSR and GIS. However, the reduction in triglyceride was more apparent in GSR (−0.48 mmol/L vs −0.13 mmol/L, −0.35, −0.58 to −0.12).

Body weight and BMI

Two trials reporting data on BMI (23) and body weight (29) showed no significant difference in the change from baseline between the two forms of glipizide (BMI: −0.2 kg/m2 vs. −0.3 kg/m2, mean difference 0.10, 95% CI: −1.19 to 1.39; Weight: −0.3 kg vs. 1.1 kg, −1.40, −2.92 to 0.12).

Subgroup analysis

All a priori hypotheses of subgroup analyses, including administration methods and treatment duration, failed to explain the heterogeneity in efficacy outcomes (interaction p > 0.10).

Adverse events

In the 14 trials of 1139 patients (18–19, 22, 24–30, 33–36) reporting hypoglycemia, 5 (19, 22, 27, 33, 35) reported no hypoglycemia events for either group. Pooling of the other 9 trials reporting relatively few hypoglycemic events (a total of 18 events with 2 vs. 16) showed that GSR had a lower rate of hypoglycemia (Peto odds ratio 0.21, 95% CI 0.08 to 0.52, Table 4).

Table 4.  Adverse events in sustained-release versus immediate-release glipzide
 No. of studiesSample sizeEvents/Total (%)Peto Odd Ratio (95% CI)Heterogeneity
Hyperglycemia(18-19, 22, 24-30, 33-36)141139  /79 (0.35%)1/60 (2.86%)0.21 (0.08 to 0.52)0.980
Gastrointestinal adverse events(24-28) 5 3651/82 (6.59%)1/83 (8.74%)0.72 (0.33 to 1.58)0.0265%
Any adverse event(18-19, 22, 24-30, 33-36)1411393/79 (6.39%)4/60 (7.32%)0.91 (0.55 to 1.48)0.1137%

Five trials of 395 patients (24–28) reported 28 gastrointestinal adverse events, including nausea, vomiting, diarrhea, epigastric pain, and abdominal distention. Of the 1139 patients (18–19, 22, 24–30, 33–36) enrolled in trials reporting adverse events, 78 reported an adverse event of any type. No statistically significant difference was found between the two drugs.

Patient compliance

Nine studies – seven RCTs and two retrospective cohort studies including 13,548 patients (19, 21, 24, 26, 28, 34–35, 37–38) – reported patient compliance. Definitions and measures of compliance differed significantly among trials: six trials (19, 21, 24, 26, 34–35) used self-reporting of missed doses; one (28) used pill counting; the two cohort studies (37–38) used the adherence index (the cumulative number of days supplied for the index drug divided by the total days in the observational period) and persistence (remaining on the previous therapy or no treatment change in the observational period). We analyzed the results of trials and cohort studies separately.

Pooled results of seven randomized trials (19, 21, 24, 26, 28, 34–35) showed that patients taking sustained-release glipizide had fewer missed doses than did those taking immediate-release gilipizide (Peto odds ratio 11.42, 95% CI 6.47 to 20.18, Table 5).

Table 5.  Patient compliance to sustained- and immediate-release glipizide
Definition of complianceDesignTreatment duration (months)No. of studiesRate of adherence (%)Risk Ratio (95% CI)
  1. *Peto odds ratio is used because of low frequency of events

Absence of missed doseRCTs(19, 21, 24, 26, 28, 34-35) 3712/28 (100%)  9/34 (70.15%)11.42 (6.47 to 20.18)*
Adherence indexCohort study(37)12145/46 (60.46%)12/46 (52.03%) 1.16 (1.02 to 1.33)
Persistence to therapyCohort study(38) 31555/618 (64.46%)204/580 (57.18%) 1.13 (1.09 to 1.16)
 Cohort study(37)12133/46 (44.37%)8/46 (35.77%) 1.24 (1.03 to 1.49)
 Cohort study(38)541202/618 (23.50%)60/580 (16.84%) 1.40 (1.29 to 1.51)

One cohort study (37) found that sustained-release glipizide increased the adherence index (60.46% vs. 52.03%, risk ratio 1.16, 95% CI 1.02 to 1.33, Table 5). Another cohort study (38), using sustained release glipizide as the control, showed that the adherence index was statistically different between the two forms (adjusted OR 0.995, 95% CI 0.991 to 0.998); however, the difference is trivial and of no clinical importance. Both cohort studies(37, 38) consistently showed that sustained-release glipizide increased the persistence to index drug therapy at 3, 12, and 54 months: 1.13 (1.09 to 1.16), 1.24 (1.03 to 1.49), and 1.40 (1.29 to 1.51) (Table 5).

Publication bias

The funnel plot and Egger's test did not identify any publication bias. The funnel plot demonstrated fairly symmetric distributions, and the Egger's tests did not show significant asymmetry (HbA1c: p = 0.559, hypoglycemia: p = 0.873).


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Declaration of interests
  9. References

In this study, we found that the reduction of HbA1c from the baseline was comparable in the two drugs analyzed; the difference in the change associated with the two drugs was minimal, and the confidence interval was narrow enough to preclude a clinically important difference. The reduction in fasting plasma glucose from the baseline appeared larger for the sustained-release glipizide. However, uncertainty remains about whether lowering fasting plasma glucose has clinical benefits. No significant difference was found in 2-hour postprandial glucose between the two drugs, and apparent uncertainty of the reduction exists.

Sustained-release glipizide decreased insulin secretion from the baseline, whereas immediate-release glipizide increased insulin secretion. This suggests that sustained-release glipizide decreases the demand on secretion of insulin while achieving comparable effects on glucose control. The mechanism underlying this effect remains unclear. One possible explanation is that sustained-release glipizide maintains a stable (with much smaller peak/trough ratio), low, and effective plasma concentration (6, 9, 10), which gently stimulates “on-demand” secretion of insulin (10). Some studies have also suggested that this phenomenon may be associated with an insulin-sensitizing effect of sustained-release glipizide (6, 9), and researchers have hypothesized that the sustained-release glipizide has a direct post-receptor effect, secondary to an improvement in glucose levels or to changes in weight or body composition (6).

The reduction of lipids level from baseline was slight in both drugs, and was comparable. The reduction in triglycerides appeared larger for the sustained-release form; apparent heterogeneity of effects existed among trials.

Results of our adverse effects analysis indicated that sustained-release glipizide led to a lower frequency of hypoglycemia. Current studies failed to show whether sustained-release glipizide reduced gastrointestinal and other adverse effects.

Both randomized trials and cohort studies suggested that the sustained release form may encourage higher compliance of administration in patients. The difference in outcomes for measuring compliance across studies, however, made it impossible to compare the consistency of magnitude of effects across cohort studies and randomized trials.

Strengths and limitations

Strengths of our analysis include our comprehensive search for studies, critical assessment of study quality, and rigorous screening of studies for eligibility and collection of data. We also developed the eligibility criteria a priori, and used a pre-specified, small number of hypotheses to explore heterogeneity. We used piloted standardized forms to collect data.

However, the included randomized trials’ lack of methodological rigor limits our ability to draw strong conclusions about the relative effects of the drugs. Inadequate reporting is common in randomized trials, resulting in difficulties in assessing methodological quality. Another apparent limitation is the underreporting of outcome data. In many of the included trials, insufficient data was reported – many reported p-values only – which led to appreciable loss of information for the analysis, and, more importantly, raises concerns about selective outcome reporting, a potentially serious consequence compromising the validity of results. Notably, many trials failed to use the change from the baseline to compare outcomes, instead comparing the values at the final follow up while ignoring the influence of baseline. This maneuver may result in misleading estimates of effects. In our study, we adjusted for this omission by using an algorithm; however, this cannot eliminate the bias inherent in these studies.

Studies included in this review generally had very limited follow up (five trials had eight weeks, while fourteen had twelve weeks), leading to insufficient assessment of the effects of the drugs on endpoints like HbA1c. Even more problematic, no trials reported outcomes important to patients. Many assumed that reduction of glucose level would result in fewer long-term macro- and micro-vascular complications. This assumption, although biologically sound, was called into doubt by a recent study (42).

Implication of study findings for clinical practice and research

Our findings suggest that sustained-release glipizide's effects on glucose control may be comparable to those of the immediate-release formulation, while inducing less insulin secretion; this combination is appealing since it may maintain islet function and sustain consistent release of insulin. In addition, simplicity of administration of sustained-release glipizide may have resulted in higher compliance in patients with type 2 diabetes mellitus. However, the short-term nature of the interventions and absence of results on long-term outcomes, along with the limited quality of the included studies, weakens the implications of these findings.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Declaration of interests
  9. References

We thank Yun Gao, Hong Zhou, Yongheng Li, and Chuan Zhang for their assistance with literature searching and selection. We also thank Mr. Peng Dong from Pfizer China for providing original articles. This study was supported by a grant from Pfizer China and The National Natural Science Foundation of China (Project No.: 71073105). The funding organizations had no role in the design and conduct of the study, analysis, or interpretation of the data, or in the preparation of and decision to submit the manuscript.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Declaration of interests
  9. References
  • 1
    IDF Clinical Guidelines Task Force. Global guideline for Type 2 diabetes. Brussels : International Diabetes Federation, 2005.
  • 2
    International Diabetes Federation. IDF DIABETES ATLAS. 4th ed. 2009. Retrieved from:
  • 3
    Fertig BJ, Simmons DA, Martin DB. Therapy for diabetes. In: National Diabetes data Group, ed. Diabetes in America. 2nd ed. Wahingtong DC : US Govt Printing Office, 1995: 951468.
  • 4
    Lebovitz HE. Glipizide: a second-generation sulfonylurea hypoglycemic agent: pharmacology, pharmacokinetics and clinical use. Pharmacotherapy 1985; 5(2): 6377.
  • 5
    Simonson DC, Kourides IA, Feinglos M, Shamoon H, Fischette CT. Efficacy, safety, and dose-response characteristics of glipizide gastrointestinal therapeutic system on glycemic control and insulin secretion in NIDDM. Results of two multicenter, randomized, placebo-controlled clinical trials. The Glipizide Gastrointestinal Therapeutic System Study Group. Diabetes Care 1997; 20(4): 597606.
  • 6
    Cefalu WT, Bell-Farrow A, Wang ZQ, Mcbride D, Dalgleish D, Terry JG. Effect of glipizide GITS on insulin sensitivity, glycemic indices, and abdominal lipid composition in NIDDM. Drug Develop Res 1998; 44(1): 17.
  • 7
    Go EH, Kyriakidou-Himonas M, Berelowitz M. Effects of glipizide GITS and glibenclamide on metabolic control, hepatic glucose production, and insulin secretion in patients with type 2 diabetes. Diabetes Metab Res Rev 2004; 20(3): 22531.
  • 8
    Feinglos M, Dailey G, Cefalu W, Osei K, Tayek J, Canovatchel W, et al . Effect on glycemic control of the addition of 2.5 mg glipizide GITS to metformin in patients with T2DM. Diabetes Res Clin Pract 2005; 68(2): 16775.
  • 9
    Berelowitz M, Fischette C, Cefalu W, Schade DS, Sutfin T, Kourides IA. Comparative efficacy of a once-daily controlled-release formulation of glipizide and immediate-release glipizide in patients with NIDDM. Diabetes Care 1994; 17(12): 14604.
  • 10
    Chung M, Kourides I, Canovatchel W, Sutfin T, Messig M, Chaiken RL. Pharmacokinetics and pharmacodynamics of extended-release glipizide GITS compared with immediate-release glipizide in patients with type II diabetes mellitus. J Clin Pharmacol 2002; 42(6): 6517.
  • 11
    Cerkoney KA, Hart LK. The relationship between the health belief model and compliance of persons with diabetes mellitus. Diabetes Care 1980; 3: 5948.
  • 12
    World Health Organization. Adherence to Long-term Therapies. Evidence for Action. Geneva : World Health Organization, 2003
  • 13
    Van Larr FA, Lucassen PLBJ, Akkermans RP. Alpha-glucosidase inhibitors for type 2 diabetes mellitus: a systematic review. Chinese Journal of Evidence-based Medicine 2006; 6(5): 33551.
  • 14
    Wells GA, Shea B, O’Connell D. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. 2009. Retrieved from:
  • 15
    Egger M, Davey Smith G, Altman DG. Systematic reviews in health care: meta-analysis in context. 2nd ed. London : BMJ Books, 2001.
  • 16
    Altman DG, Bland JM. Interaction revisited: the difference between two estimates. BMJ 2003; 326: 219.
  • 17
    Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics 1994; 50: 1088101.
  • 18
    Zhang X. The effect of extended-release glipizide for type 2 diabetes mellitus. Fujian Medical Journal 2000; 22(3): 234. (In Chinese)
  • 19
    Wang P, Yao C, Wang W. The comparison of efficacy between glipizide GITS and GIR for type 2 diabetes mellitus. China pharmacy 2001; 12(4): 230. (In Chinese)
  • 20
    Zeng W, Shen H. The extended-release glipizide for type 2 diabetes mellitus. Journal of Shanghai Second Medical University 2001; 21(1): 845. (In Chinese)
  • 21
    Chen B, Cai H, Luo M. Randomized clinical trial of Glucotrol XL, a release-controlled glipizide, in treatment for type 2 diabetes. Acta Acadimiae Medicinae Militaris Tertiae 2003; 25(2): 1645. (In Chinese)
  • 22
    Guo X, Huang Y, Liang D. Clinical effects of glipizide gastrointestinal therapeutic system in treatment of secondary failure to sulfonylurea in noninsulin dependent diabetes mellitus. Chinese Journal of Clinical Pharmacy 2004; 13(3): 13941. (In Chinese)
  • 23
    Li C, Xue Y, Gao F. The effects on the secretion of islet and sensitivity of insulin of sustained-release glipizide for type 2 diabetes mellitus. Guangdong Medical Journal 2004; 25(6): 71920. (In Chinese)
  • 24
    Shen L, Chen K, Xie Y. Sustained-release glipizide for the treatment of type 2 diabetes mellitus among 30 patients. Herald of Medicine 2004; 23(11): 8289. (In Chinese)
  • 25
    Ji L. The Clinical effects of extended-release glipizide for type 2 diabetes mellitus. Modern practical Medicine 2005; 17(8): 4923. (In Chinese)
  • 26
    Wang J, Jiang L, Cai J. The pharmacoeconomic study of different forms of glipizide for type 2 diabetes mellitus. Chinese Journal of Hospital Pharmacy 2005; 25(9): 8667. (In Chinese)
  • 27
    Yang X. Influence of blood glucose and blood lipid of dyslipidosis associated with type 2 diabetes mellitus patients with glipizide GITS. Chinese General Practice 2005; 8(4): 27980. (In Chinese)
  • 28
    Hsieh SH, Lin JD, Cheng HY, Ho C, Liou MJ. Sustained-release versus immediate-release glipizide for treatment of type 2 diabetes mellitus in Chinese patients: A randomized, double-blind, double-dummy, parallel-group, 12-week clinical study. Clin Ther 2006; 28(9): 131826.
  • 29
    Yao M, Shang Y, Xu Q. The effects on sensitivity to insuline of glipizide XL for type 2 diabetes. Sichuan Medical Journal 2006; 27(7): 7134. (In Chinese)
  • 30
    Zhao X, Zhou Y, Yang J. Clinical study of controlled-release formula of glipizide in the treatment of coronary artery disease complicated with type 2 diabetes mellitus. Journal of Capital University of Medical Sciences 2006; 27(1): 7880. (In Chinese)
  • 31
    Wu Y. the comparison of efficacy between glipizide GITS and immediate-release among the old patients with type 2 diabetes mellitus. Journal of Practical Diagnosis and Therapy 2007; 21(2): 135. (In Chinese)
  • 32
    Zhang W, Wu H. the efficacy between glipizide GITS and glipizide for type 2 diabetes mellitus. Chinese Community Doctors 2007; 21(5): 26. (In Chinese)
  • 33
    Cai M, Yang S, Qi Y. the comparison of efficacy between glipizide GITS and immediate-release glipizide for type 2 diabetes mellitus. Chongqing Medical Journal 2008; 29(5): 52930. (In Chinese)
  • 34
    Deng LY. Glipizide GITS and immediate-release glipizide in treatment of type 2 diabetes mellitus. China Modern Medicine 2010; 17(21): 734. (In Chinese)
  • 35
    Yao P, Zhao LX, Wang Z. The clinical study on treating type 2 diabetes mellitus with glipizide. Clinical Journal of Chinese Medicine 2010; 2(18): 267. (In Chinese)
  • 36
    Zhang HF, Qi JH. Efficacy and safety of sustained-release glipizide for type 2 diabetes mellitus. China Practical Medicine 2011; 6(5): 1634. (In Chinese)
  • 37
    Dezii CM, Kawabata H, Tran M. Effects of once-daily and twice-daily dosing on adherence with prescribed glipizide oral therapy for type 2 diabetes. South Med J 2002; 95(1): 6871.
  • 38
    Harley CR, Mishra A, Cantor SE, Nelson M, Walker AM. A ‘real-world’ analysis of persistence on and adherence to glipizide GITS, glipizide IR and glibenclamide in managed care among patients with type 2 diabetes mellitus. Clin Drug Invest 2002; 22(9): 57584.
  • 39
    World Health Organization. Definition, diagnosis and classification of diabetes mellitus and its complications – report of a WHO consultation. Part 1: diagnosis and classification of diabetes mellitus. WHO/NCD/NCS/99.2. Geneva : World Health Organization, 1999.
  • 40
    Gavin JR, Alberti KGMM, Davidson MB. Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care 1997; 20: 118397.
  • 41
    World Health Organization. Diabetes Mellitus: Report of a WHO Study Group. Technical Report Series No. 727. Geneva : World Health Organization, 1985.
  • 42
    Kelly TN, Bazzano LA, Fonseca VA, Thethi TK, Reynolds K, He J. Systematic review: glucose control and cardiovascular disease in type 2 diabetes. Ann Intern Med 2009; 151: 394403.