Abstract
Objectives
This study aimed to assess if metformin had any associations with the prevalence of obstructive sleep apnea in an adult type 2 diabetes population in the Midwest.
Hypothesis
Use of metformin is associated with decreased prevalence of obstructive sleep apnea in a adult type 2 diabetes population.
Methods
A retrospective secondary database analysis was carried out with metformin use by patients with type 2 diabetes as the primary variable of interest and obstructive sleep apnea status as the primary outcome. A sample population of 9,853 type 2 diabetes patients with one year of follow-up was used. Other variables that were analyzed included age, gender, race, hypertension, Congestive Heart Failure, Hemoglobin A1c (HbA1c), and Body Mass Index. A p-value of <0.01 was considered significant.
Results
Metformin usage was not significantly associated with obstructive sleep apnea prevalence (Odds Ratio: 1.17, Confidence Interval: 1.00–1.36, p = 0.049), but trended in the direction where metformin usage was associated with having obstructive sleep apnea. Lower HbA1c was found to be significantly associated with lower prevalence of obstructive sleep apnea (p <0.001). The rest of the variables followed previously published associations.
Conclusions
Metformin therapy may improve sleep quality, but it may not be through methods that reduce the likelihood of developing obstructive sleep apnea. Future studies that can prove causation about this association should be considered.
Keywords: Obstructive sleep apnea, Metformin, Type 2 diabetes, Insulin resistance
INTRODUCTION
The relationship between insulin resistance and sleep apnea remains unclear. There have been multiple population studies that have shown that they are correlated, however no one has been able to discern the exact nature of the relationship [1–5]. Treatment of obstructive sleep apnea using continuous positive airway pressure appears to improve diabetes outcomes (including insulin resistance) modestly, suggesting that methods of treating sleep apnea could be used for treating type 2 diabetes [1,6–8].
Treatment of type 2 diabetes improving obstructive sleep apnea outcomes is a possibility that remains largely uninvestigated. There are theories that diabetic neuropathy might contribute to increased collapsibility of the airway. [1,4,5,9]. A study discovered that metformin treatment, the first-line drug used for type 2 diabetes, reduced sleep apnea scores for central sleep apnea in a non-obese rat model [10,11]. Another study found that metformin use was associated with longer sleep duration and better sleep quality in humans [12]. However, there have been no studies on the effects of metformin on obstructive sleep apnea, which is more clinically relevant to type 2 diabetes patients compared to central sleep apnea in humans. There is also only sparse data with regards to diabetic variables that affect obstructive sleep apnea, and no studies that observe those variables related to obstructive sleep apnea specifically in type 2 diabetic patients. This study aims to provide insight to that relationship while taking into account the available confounding variables.
MATERIALS AND METHODS
A retrospective study examining use of metformin in type 2 diabetes patients with and without obstructive sleep apnea was carried out. The Medical College of Wisconsin Human Subjects Review Board approved this medical review using de-identified patient data. Patient data was extracted from the Clinical Research Data Warehouse, which contains medical record data from all patients served by Froedert Hospital and the Medical College of Wisconsin. Data was analyzed for clinical encounters from January 1st 2006 to December 31st 2013, with patients having various follow-up times. A random subset of 10,000 patients was extracted from the group of patients with at least one year of follow up after their first diagnosis of type 2 diabetes defined by ICD-9 codes 250.x0 and 250.x2. Patients under 18 or over 88 years of age were excluded. After the exclusion, 9,853 patients remained. The outcome of interest in the study was prevalence of obstructive sleep apnea (defined as an ICD 9 code of 327.23). The key independent variable of interest was the self-reported metformin use (defined as any drug in class2725 Biguanides). Patients were asked what drugs they were taking and if metformin use was reported any time over the past year it was recorded in the electronic medical record. Other variable data was found through cross comparisons with physician reports and billing information. The association between obstructive sleep apnea was analyzed in unadjusted and adjusted for other confounding variables analyses. Descriptive statistics were used to summarize sample characteristics. Potential factors associated with obstructive sleep apnea status included age, sex, race, hypertension, congestive heart failure (CHF), Hemoglobin A1c (Hba1c), and body mass index (BMI). Age, HbA1c and BMI were initially analyzed as continuous variables. A Mann-Whitney test was used to compare distributions of continuous variables between sleep apnea groups. Later, HbA1c and BMI were considered as categorical variables for ease of analysis. A chi square test was used for bivariate analysis of each categorical variable by obstructive sleep apnea status. A backwards variable selection with a statistical significance alpha level of 0.05 was then carried out on the full model with all potential predictors, which removed race and age. Finally, a logistic regression of obstructive sleep apnea on sex, hypertension, CHF, HbA1c, and BMI was fitted. An alpha level of 0.01 was used to declare statistical significance. SAS on Demand Enterprise Guide 4.3 (SAS Institute, Cary, NC) was used to perform all statistical analysis. The alpha value of 0.01 was chosen to limit the number of false discovery findings often associated with large sample sizes.
RESULTS AND DISCUSSION
Descriptive characteristics of the patients with diabetes
The descriptive analysis is summarized in Table 1. A majority of the population with diabetes did not have sleep apnea (n= 8949, 90.8%). There were significant differences in sleep apnea rates in multiple demographic characteristics. The obstructive sleep apnea rate was 4% higher for diabetes patients on metformin (p <0.0001) and was 3% higher for males (p <0.0001). There was no significant difference in the prevalence of sleep apnea between race groups (p = 0.056). Alcohol status was not significantly associated with sleep apnea rate for those patients whose information was known, though the patients with unknown information were significantly less likely to have obstructive sleep apnea (p<0.0001). There was a 15% higher rate of sleep apnea for patients with Congestive Heart Failure (p <0.0001). Hypertension was associated with an 8% increase of sleep apnea (p<0.0001). BMI showed a positive association with sleep apnea (p<0.0001). HbA1C showed a negative association with sleep apnea (p<0.0001).
Table 1.
Characteristics of Subjects with and without Apnea.
| Characteristic | No Apnea Number (%) | Apnea Number (%) | P Value |
|---|---|---|---|
| Metformin Use | <0.0001 | ||
| Yes | 3053 (88) | 417 (12) | |
| No | 5896 (92) | 487 (8) | |
| Sex | <0.0001 | ||
| Male | 4240 (89) | 514 (11) | |
| Female | 4709 (92) | 390 (8) | |
| Race | 0.056 | ||
| White | 6396 (91) | 655 (9) | |
| Black | 1866 (90) | 200 (10) | |
| Other | 663 (93) | 48 (7) | |
| Alcohol Use | <0.0001 | ||
| Yes | 2929 (89) | 363 (11) | |
| No | 3216 (89) | 395 (11) | |
| Unknown | 2804 (95) | 146 (5) | |
| Congestive Heart Failure | <0.0001 | ||
| Yes | 611 (77) | 182 (23) | |
| No | 8339 (92) | 722 (8) | |
| Hypertension | <0.0001 | ||
| Yes | 5496 (88) | 759 (12) | |
| No | 3454 (96) | 145 (4) | |
| BMI | <0.0001 | ||
| <30 | 2889 (96) | 109 (4) | |
| 30–35 | 1709 (91) | 170 (9) | |
| 35–40 | 1086 (86) | 181 (14) | |
| >40 | 1055 (75) | 361 (25) | |
| Missing | 2210 (96) | 83 (4) | |
| HbA1C | <0.0001 | ||
| <7 | 2708 (88) | 360 (12) | |
| 7–8.5 | 1489 (90) | 157 (10) | |
| >8.5 | 938 (92) | 89 (9) | |
| Missing | 3814 (93) | 298 (7) |
Multivariable analysis
The regression analysis results are summarized in Table 2 and Table 3. There was not a significant association between metformin use and obstructive sleep apnea (OR 1.17, 95%, CI 1.00, 1.36, p = 0.049). Among potential confounders, females had half the odds of having obstructive sleep apnea compared to males, which has been supported in previous studies. Presence of CHF was positively associated with having obstructive sleep apnea, with odds two and a half times higher than those who did not have CHF. Similarly, having hypertension was associated with twice the odds of having obstructive sleep apnea. A higher BMI category was associated with greater odds of having obstructive sleep apnea. A lower HbA1c was associated with greater odds of obstructive sleep apnea. There was a significant interaction between alcohol and hypertension as well as alcohol and HbA1c which were included in the final model. This study demonstrated that metformin usage was not associated with obstructive sleep apnea prevalence. Gender, alcohol use, hypertension, CHF, BMI and Hba1c levels were all significant variables associated with obstructive sleep apnea that were controlled for in the model. The associations of the variables agreed with previous research, with the exception of Hba1c levels [13,14] Lower Hba1c levels were associated with increased sleep apnea prevalence.
Table 2.
Logistic Regression Analysis of Key Study Variables.
| Variable | Level_Var1 | Level_Var2 | Odds Ratio | Lower CI | Upper CI | P Value |
|---|---|---|---|---|---|---|
| Intercept | 0.03 | 0.02 | 0.05 | <.001 | ||
| Metformin | Yes | 1.17 | 1.00 | 1.36 | 0.049 | |
| Hypertension | Yes | 2.25 | 1.61 | 3.14 | <.001 | |
| Sex | Female | 0.49 | 0.42 | 0.57 | <.001 | |
| Alcohol Use | Yes | 1.48 | 0.94 | 2.32 | 0.091 | |
| Alcohol Use | Missing | 0.26 | 0.15 | 0.46 | <.001 | |
| Congestive Heart Failure | Yes | 2.66 | 2.18 | 3.26 | <.001 | |
| HbA1C | [7 – 8.5] | 0.69 | 0.50 | 0.95 | 0.021 | |
| HbA1C | >= 8.5 | 0.56 | 0.39 | 0.81 | 0.002 | |
| HbA1C | Missing | 0.55 | 0.42 | 0.72 | <.001 | |
| BMI | [30 – 35] | 2.67 | 2.07 | 3.44 | <.001 | |
| BMI | [35 – 40] | 4.75 | 3.68 | 6.13 | <.001 | |
| BMI | >= 40 | 10.60 | 8.35 | 13.46 | <.001 | |
| BMI | Missing | 1.42 | 1.04 | 1.95 | 0.030 | |
| Alcohol*Hypertension | Yes | Yes | 0.55 | 0.35 | 0.85 | 0.008 |
| Alcohol *Hypertension | Missing | Yes | 1.77 | 1.05 | 2.98 | 0.033 |
| Alcohol *HbA1C | Yes | [7 – 8.5] | 1.05 | 0.67 | 1.65 | 0.826 |
| Alcohol *HbA1C | Yes | >= 8.5 | 0.85 | 0.49 | 1.47 | 0.558 |
| Alcohol *HbA1C | Yes | Missing | 1.30 | 0.88 | 1.92 | 0.189 |
| Alcohol *HbA1C | Missing | [7 – 8.5] | 0.93 | 0.45 | 1.91 | 0.840 |
| Alcohol *HbA1C | Missing | >= 8.5 | 1.08 | 0.43 | 2.72 | 0.873 |
| Alcohol *HbA1C | Missing | Missing | 3.13 | 1.90 | 5.16 | <.001 |
Table 3.
Effects of Key Variables on Degrees of Freedom in Multivariable Model.
| Effect | Degrees of Freedom | P Value |
|---|---|---|
| Metformin | 1 | 0.049 |
| Hypertension | 1 | <.001 |
| Sex | 1 | <.001 |
| Alcohol Use | 2 | <.001 |
| Congestive Heart Failure | 1 | <.001 |
| HbA1C | 3 | <.001 |
| BMI | 4 | <.001 |
| Alcohol*Hypertension | 2 | <.001 |
| Alcohol*HbA1C | 6 | <.001 |
The direction of the association of metformin and obstructive sleep apnea maybe due to the population differences in the metformin and non-metformin using groups. Patients who have started metformin therapy are typically at a more advanced stage of type 2 diabetes than those who control their blood sugar level with diet and exercise. HbA1c levels were also generally lower in the obstructive apnea group. The population using metformin might have lower HbA1c because they are using a drug with the purpose to lower their HbA1c levels. An analysis of metformin compared to other blood glucose controlling drugs such as insulin, Sulfonyl ureas, and other oral anti-glycemic medications could be used to discern if the explanation is in fact what is associated with that trend.
Metformin therapy may still have the potential to improve sleep quality, but it may not be through methods that would reduce the likelihood of developing obstructive sleep apnea. The study had several limitations. As with all secondary database analyses, causation cannot be proven and variables that would have been useful to include in the model such as neck circumference and sleep scores were unavailable. Only the diagnosis of sleep apnea was available and there was no information on its severity or whether the patient was in the process of treatment (such as CPAP) which would introduce confounders. There were groups with some unknown or missing data as well, such as alcohol use, BMI, and HbA1C. The diagnosis of obstructive sleep apnea was a patient self-reported measure as well. The dataset may also include a few miscoded diagnoses of Type I diabetes, secondary diabetes, and gestational diabetes which were unadjusted for in the model. Researchers may want to look into mechanisms of metformin or other type 2 diabetes drugs in relationship to mechanisms contributing to obstructive sleep apnea for future studies. Clinicians should not be too alarmed by these results, but should still be aware that diabetes and sleep apnea are intertwined when consulting or treating their patients for either disease.
CONCLUSION
Metformin use was not associated with reduced obstructive sleep apnea prevalence. Other markers for obstructive sleep apnea followed previous research with the exception of HbA1c levels. This was the first study on the topic that has used a large number of subjects. Future studies should consider doing a randomized control trial to investigate causation, as well as test other blood glucose lowering drugs’ effects on obstructive sleep apnea. This descriptive study of the association of metformin and obstructive sleep apnea is clinically important because obesity in type 2 diabetes contributes significantly to OSA which in turn increases the risks of cardiovascular disease in these patients. Metformin is prescribed to improve diabetes control but may not help significantly with OSA management. Other interventions such as continuous positive airway pressure may be needed to manage this complication.
Acknowledgments
Lisa Rein, Sergey Tarima, and John Meurer were partly funded by the MCW Advancing a Healthier Wisconsin endowment for population-based diabetes research and education.
References
- 1.Tahrani AA, Ali A, Stevens MJ. Obstructive sleep apnoea and diabetes: an update. Curr Opin Pulm Med. 2013;19:631–638. doi: 10.1097/MCP.0b013e3283659da5. [DOI] [PubMed] [Google Scholar]
- 2.Chasens ER. Obstructive sleep apnea, daytime sleepiness, and type 2 diabetes. Diabetes Educ. 2007;33:475–482. doi: 10.1177/0145721707301492. [DOI] [PubMed] [Google Scholar]
- 3.Shaw JE, Punjabi NM, Wilding JP, Alberti KG, Zimmet PZ International Diabetes Federation Taskforce on Epidemiology and Prevention. Sleep-disordered breathing and type 2 diabetes: a report from the International Diabetes Federation Taskforce on Epidemiology and Prevention. Diabetes Res Clin Pract. 2008;81:2–12. doi: 10.1016/j.diabres.2008.04.025. [DOI] [PubMed] [Google Scholar]
- 4.Nannapaneni S, Ramar K, Surani S. Effect of obstructive sleep apnea on type 2 diabetes mellitus: A comprehensive literature review. World J Diabetes. 2013;4:238–244. doi: 10.4239/wjd.v4.i6.238. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Aurora RN, Punjabi NM. Obstructive sleep apnoea and type 2 diabetes mellitus: a bidirectional association. Lancet Respir Med. 2013;1:329–338. doi: 10.1016/S2213-2600(13)70039-0. [DOI] [PubMed] [Google Scholar]
- 6.Iftikhar IH, Khan MF, Das A, Magalang UJ. Meta-analysis: Continuous Positive Airway Pressure Improves Insulin Resistance in Patients with Sleep Apnea without Diabetes. Ann Am Thorac Soc. 2013;10:115–120. doi: 10.1513/AnnalsATS.201209-081OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Zhong W, Tang YG, Zhao X, Go FY, Harper RM, Hui H. Treating Obstructive Sleep Apnea With Continuous Positive Airway Pressure Benefits Type 2 Diabetes Management. Pancreas. 2014;43:325–330. doi: 10.1097/MPA.0000000000000083. [DOI] [PubMed] [Google Scholar]
- 8.Chasens ER, Strollo PJ., Jr Treatment of obstructive sleep apnea on insulin resistance: not an “anti-sugar pill. Ann Am Thorac Soc. 2013;10:150–151. doi: 10.1513/AnnalsATS.201302-031ED. [DOI] [PubMed] [Google Scholar]
- 9.Fujihara K, Kodama S, Horikawa C, Yoshizawa S, Sugawara A, Hirasawa R, et al. The Relationship between Diabetic Neuropathy and Sleep Apnea Syndrome: A Meta-Analysis. Sleep Disord. 2013;2013:150371. doi: 10.1155/2013/150371. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Ramadan W, Dewasmes G, Petitjean M, Wiernsperger N, Delanaud S, Geloen A, et al. Sleep apnea is induced by a high-fat diet and reversed and prevented by metformin in non-obese rats. Obesity (Silver Spring) 2007;15:1409–1418. doi: 10.1038/oby.2007.169. [DOI] [PubMed] [Google Scholar]
- 11.Ramadan W, Petitjean M, Loos N, Geloen A, Vardon G, Delanaud S, et al. Effect of high-fat diet and metformin treatment on ventilation and sleep apnea in non-obese rats. Respir Physiol Neurobiol. 2006;150:52–65. doi: 10.1016/j.resp.2005.02.011. [DOI] [PubMed] [Google Scholar]
- 12.Kajbaf F, Fendri S, Basille-Fantinato A, Diouf M, Rose D, Jounieaux V, Lalau JD. The relationship between metformin therapy and sleep quantity and quality in patients with Type 2 diabetes referred for potential sleep disorders. Diabet Med. 2014;31:577–580. doi: 10.1111/dme.12362. [DOI] [PubMed] [Google Scholar]
- 13.Ralls FM, Grigg-Damberger M. Roles of gender, age, race/ethnicity, and residential socioeconomics in obstructive sleep apnea syndromes. Curr Opin Pulm Med. 2012;18:568–573. doi: 10.1097/MCP.0b013e328358be05. [DOI] [PubMed] [Google Scholar]
- 14.Deegan PC, Mc Nicholas WT. Pathophysiology of obstructive sleep apnoea. Eur Respir J. 1995;8:1161–1178. doi: 10.1183/09031936.95.08071161. [DOI] [PubMed] [Google Scholar]