Continuous or Intermittent Monitoring of Glucose in Interstitial Fluid*Medical Policy: 01.01.03 This policy applies to all products unless specific contract limitations, exclusions or exceptions apply. Please refer to the member's coverage manual for benefit availability. Managed care guidelines related to referral authorization, and precertification of inpatient hospitalization, home health, home infusion and hospice services apply. The use of blood glucose monitors in the home has greatly impacted the management of diabetes. Several clinical trials have demonstrated that decreases in diabetes complications are associated with tight glucose control, defined as HbA1c measurement of less than 7%. Recently, measurements of glucose obtained from the interstitial fluid have been utilized as a technique for automatically measuring glucose values throughout the day. Although the time interval at which interstitial glucose is measured ranges from every 5 to 10 minutes, all such monitoring devices are commonly referred to as continuous glucose monitors. These devices do not eliminate the need for finger stick blood glucose measurements. Two to four times a day calibration must be done with finger stick blood glucose readings and each interstitial fluid glucose reading must be validated by finger stick blood glucose readings before any adjustments in insulin or other interventions are made. This is true even if the device is associated with an insulin pump. Examples of these devices include;
Additional devices now have FDA approval:
In evaluating the continuous glucose monitoring systems, it is important to recognize that they may be used intermittently, e.g., time periods of 72 hours, or continuously. Prior approval is recommended. Submit a prior approval now. Intermittent monitoring of glucose levels in the interstitial fluid as a technique of monitoring type 1 diabetes may be considered medically necessary for up to three days at a time for patients who:
When one of the following criteria is met:
Intermittent monitoring of glucose levels in the interstitial fluid as a technique of monitoring suspected non-diabetic hypoglycemia, such as may occur with Nesidioblastosis (islet cell dysmaturation syndrome) and insulinoma may be considered medically necessary for up to three days at a time. Written documentation from the medical record specifying the medical necessity, according to the criteria above, is required. Since time is needed to determine the efficacy of these treatment modifications, it is anticipated that when the above medical necessity criteria are met, the device could be used up to four times in 12 months. Medical Director review is required if requests are received to use this device more than four times in 12 months. Devices intended to monitor continuous interstitial glucose monitoring beyond three days (longer than 72 hrs) as an adjunct to standard care is considered medically necessary for patients with type 1 diabetes when the following criteria are met:
Written documentation from the medical record specifying the medical necessity, according to the criteria above, is required. Medical Director review is required if requests are received to use this device for patients under the age of 18 years. Rationale: Data presented to the U.S. Food and Drug Administration (FDA) advisory committee meeting consisted of studies validating the correlation between the measurements of glucose in interstitial fluid with the blood glucose measurements made with home monitoring devices. While the individual values between the two may vary, in general, the panel found that the overall trends in glucose levels detected by frequent measurements produced potentially clinically important information. However, there were no clinical data presented regarding improvements in HbA1c measurements or a decreasing incidence of hypoglycemic episodes in those whose antidiabetic medications were managed based on more frequent readings of interstitial fluid glucose. However, members of the advisory panel felt that more frequent measurements should extrapolate to improved diabetic management. For example, prior studies have shown that HbA1c levels are lowest among patients who have the highest frequency of daily blood glucose measurements. The key clinical outcomes regarding the clinical utility of interstitial measurements of glucose, using either the Continuous Glucose Monitoring Systems or the GlucoWatch G2 Biographer, relates to their ability to provide either additional information on glucose levels leading to improved glucose control, or to improve the morbidity and mortality associated with clinically significant severe and acute hypoglycemic or hyperglycemic events. Because diabetes control encompasses numerous variables including the diabetes regimen and patient self-management, randomized controlled trials are important to isolate the contribution of interstitial glucose measurements to the overall diabetes management. This policy is based on a 2003 TEC Assessment which reviewed the published controlled trials and offered the following discussion. GlucoWatch Biographer Chase and colleagues reported the results of a trial of 40 children with poorly controlled type 1 diabetes (HbA1c >8) who were randomized to diabetic management with or without glucose monitoring with a GlucoWatch device. Conventional glucose monitoring was performed 4 times daily in both groups. Those randomized to the treatment group were asked to wear the device 4 times per week for three months. After three months, all patients received Biographers and were followed up for six months. HbA1c values were determined at baseline and after one, three, six, and nine months. The median HbA1c level dropped from 8.9% to 8.4% in the treatment group, while in the control group the HbA1c increased from 8.6% to 9%. While this difference was statistically significant, it should be noted that the worsening of HbA1c in the control group was nearly as large in magnitude as the improvement in HbA1c seen in the Biographer group. There was no significant improvement in “fear of hypoglycemia” or quality of life between the two groups. In a second observational phase of the trial, all subjects were provided Biographer devices and observed over an additional six months. During this phase, the Biographer group maintained median HbA1c at 8.5%, and the control group improved median HbA1c to 8.6%, which was their original level. It was noted that the frequency of use of the Biographer declined over the course of the first phase of the study, and this may be why the Biographer group did not show further improvement in HbA1c over the subsequent 6 months of use. Baseline characteristics of the two randomized study groups were reported to be without statistically significant differences. However, the baseline median HbA1c levels for these two groups were different by 0.3%, which is almost as large as the 0.4%–0.5% change observed within groups after provision of the Biographer, and this difference may have clinical significance. Also at baseline, slightly more patients in the control group used insulin pumps or received three or more insulin injections per day compared with the Biographer group, which had slightly more subjects receiving only two insulin injections per day. It is unknown whether these slight imbalances were a result of the small sample or whether there were any problems with randomization. The authors do not discuss whether such differences might have influenced the observed results, but additional analyses adjusting for differences in potentially confounding baseline characteristics and exploring whether outliers could have influenced the results would be of interest. In addition, it is unclear whether subjects in the Biographer group received more frequent or more intense contact with physicians and the diabetes clinic. Biographer subjects were required to visit the clinic each week to download Biographer data; whereas the control group was able to fax back conventional fingerstick glucose meter data. This process may have provided more in-person opportunity for medical input in the Biographer group. Interpretation of this study’s results should also take into consideration the observation that HbA1c levels may fluctuate over time, even without intervention, and variations of up to 1% may be observed clinically in the pediatric population. In this study, the control group’s HbA1c got worse during the intervention study, which partially contributed to the statistically significant difference between groups. Improvement in HbA1c has been observed in control groups in multiple other studies, most likely as a result of study effects (Hawthorne effect) in which participants in a trial achieve better compliance when results are being monitored. It is unclear why the control group got worse in this study, and this raises concerns over the reproducibility of the study. Interpretation of the clinical significance of reducing HbA1c by 0.5% has been explored and both magnitude and durability of the improvement are important factors to consider. Eastman and colleagues presented an abstract of a decision analysis model based on the above study and reported that “the model predicts that treating 100 subjects under Biographer-guided standard care, if maintained for the life of the cohort, would prevent 20 cases of proliferative retinopathy, four cases of macular edema, six cases of blindness, 12 cases of clinical albuminuria, eight cases of end-stage renal disease, six cases of neuropathy and one amputation.” However, this model makes a variety of assumptions regarding the durability of the improvement. In summary, Chase and colleagues conducted a small, randomized controlled trial and reported a small but statistically significant difference in the median HbA1c levels between groups after three months. However, the relatively small magnitude of incremental improvement in HbA1c levels needs to be interpreted in the context of potentially different baseline statistics between subjects in the two groups, potential study effects (Hawthorne effect) in the Biographer group in this unblinded trial, and potential influences of receiving more intense medical attention in the Biographer group. It would be very helpful to see the results of this trial confirmed by another larger, multicenter randomized controlled trial and to have further studies explore the durability of HbA1c improvements over time. Continuous Glucose Monitoring Systems (CGMS) Results of four randomized trials have been reported. The largest of them, which enrolled 128 adult patients with type 1 diabetes, is available in abstract only. Among the 109 patients completing the three-month trial (the dropout rate was 15%), there was no statistically significant difference in HbA1c levels. Mean HbA1c levels in both the control and study groups declined from 9% at baseline to 8.3% at three months. Similarly, in another randomized study of 75 patients, there was no statistical difference in HbA1c levels after the three-month intervention. The other randomized studies included only 11 and 27 patients, respectively. In 2004, Tanenberg and colleagues reported on a study of 128 patients randomized to insulin therapy adjustments using data from either the CGMS or self-monitoring of blood glucose (SMBG) using a home blood glucose monitor over a 12-week period. At 12 weeks, HbA1c levels and hyperglycemic event frequency and duration did not differ with any statistical significance in the treatment groups. However, at 12 weeks, events of hypoglycemia (glucose < or = 60 mg/dL) were found to be significantly shorter in the CGMS group than in the SMBG group (49.4 +/- 40.8 vs. 81.0 +/- 61.1 minutes per event, p = .009). The authors concluded that durations of hypoglycemia can be further reduced by adjusting insulin therapy with data from the CGMS rather than using SMBG data alone. Nevertheless, the biochemically defined measurements of hypoglycemia (without accompanying evidence of symptoms or a clinically significant hypoglycemic event) are not compelling outcomes. The clinical significance of these test results has not been established, i.e., there is insufficient evidence showing the link between increased duration of asymptomatic hypoglycemia and subsequent clinical outcomes. 2006 — 2007 Update Additional studies continue to evaluate continuous glucose monitoring systems. Lagarde and colleagues found a slight improvement in HbA1c levels using CGMS compared to controls in children with type 1 diabetes. However, the difference did not reach statistical significance (p = 0.13). In a European study using a cross-over design, Deiss and colleagues reported that CGMS did not decisively influence glycemic control of the total study cohort of children and adolescents with type 1 diabetes. They suggested that more frequent use of CGMS at shorter intervals may be of greater value. A recent review in The Medical Letter on Drugs and Therapeutics raised questions about the accuracy of these systems. Garg, et al. reported in 2006 that in 91 patients with diabetes (75 were type 1) real-time continuous glucose monitoring was able to reduce glycemic excursions by reducing hyperglycemia without increasing the risk of hypoglycemia. They also indicated that this type of monitoring may reduce long-term complications of diabetes. Recently, Deiss, et al. reported on a three-month study of 81 children and 81 adults with stable type 1 diabetes who had HbA1c levels of 8.1% or greater. Patients were randomized to continuous real-time monitoring, continuous monitoring for three days every two weeks, or self-monitoring of blood glucose. At three months, 50% of patients with continuous real-time monitoring had a decrease in HbA1c of at least 1% compared to 37% of those with intermittent continuous monitoring, and 15 % of controls. These results suggest that continuous glucose monitoring may have potential for improving control in patients with diabetes; however, as the authors note, additional work is needed to determine long-term efficacy, clinical feasibility in patients with varying levels of glycemic control, and effect on rates of hypoglycemia. April 2008 Update The policy was updated with a literature review using MEDLINE from January 2007 through February 2008. No publications were identified that present results from randomized trials that show an impact of long-term continuous glucose monitoring on relevant patient outcomes. Recent publications continue to report results on case series and often do not clearly link to patient outcomes. Guillod reported on a retrospective study that described findings from a group of 88 patients with type 1 diabetes who underwent a CGMS exam. The prevalence of nocturnal hypoglycemia (NH) was 67% (32% of them unsuspected). A measured hypoglycemia at bedtime (22–24 hr) had a sensitivity of 37% to detect NH, while a single measure 4 mmol/l or less at 03-hour had a sensitivity of 43%. In this study, NH measurements were not associated with morning hyperglycemias but with morning hypoglycemias. After 6–9 months, suspicions of NH decreased from 60% to 14% (p<0.001). The authors concluded that NH was highly prevalent and often undetected. Self-monitoring blood glucose at bedtime, which detected hypoglycemia, had sensitivity almost equal to that of 03-hour and should be preferred because it is easier to perform. Tubiana-Rufi reported on an uncontrolled study of 182 patients (children and adults) with poorly controlled type 1 diabetes. Using the Guardian RT system, which the authors indicated required 3 calibrations a day, resulted in improvement in HbA1c levels over 3 months. The DirecNet Study Group reported results of another non-comparative study of 30 patients with type 1 diabetes who used an insulin pump with the FreeStyle Navigator CGM system for 13 weeks. During this time, the mean HbA1c levels improved from 7.1% to 6.8% and the percentage of glucose values between 71 and 180 mg/dl increased from 52% to 60%. Two patients had severe skin reactions related to the sensor mount adhesive. Wilson and colleagues, as part of the Diabetes Research in Children Network (DirecNet), evaluated the accuracy and precision of the FreeStyle Navigator CGMS in 30 children with type 1 diabetes (mean age 11.2 years). The Navigator glucose values were compared with reference serum glucose values of blood samples obtained in an inpatient clinical research center and measured in a central laboratory and in an outpatient setting with a FreeStyle meter. Median absolute difference (AD) and median relative absolute difference (RAD) were computed for sensor-reference and sensor-sensor pairs. The median AD and RAD were 17 mg/dl and 12%, respectively, for 1,811 inpatient sensor-reference pairs, and 20 mg/dl and 14%, respectively, for 8,639 outpatient pairs. The median RAD between two simultaneous Navigator measurements (n = 1,971) was 13%. Ninety-one percent of sensors in the inpatient setting and 81% of sensors in the outpatient setting had a median RAD of 20% or less. The authors concluded that the Navigator's accuracy does not yet approach the accuracy of current-generation home glucose meters, but it is sufficient to believe that the device has the potential to be an important adjunct to treatment of youth with type 1 diabetes. Several authors note that these results provide a compelling rationale for conducting a randomized controlled trial (RCT) of use of continuous glucose monitoring in type 1 diabetes. Recent advances in technology now allow linkage between the CGM device and an insulin pump. Halvorson reported on an uncontrolled pilot trail of 10 children with type 1 diabetes. The small size and lack of control group limit the ability to draw any conclusions from this study. Publications are also beginning to report on early trials of use of these devices in patients with type 2 diabetes. Wolpert discussed the skills needed for diabetes management using real-time monitoring and commented specifically on the role of calibration as well as understanding the lag between capillary and interstitial glucose levels. Given the lack of scientific data about the impact of CGM on clinical outcomes, this is considered investigational. October 2008 Update The policy was updated in October 2008 with a literature search using MEDLINE. A recent systematic review of randomized studies identified 7 studies with 335 patients that fulfilled their inclusion criteria. Study duration varied from 12 to 24 weeks. This review concluded that compared with self-monitoring, CGMS was associated with a non-significant reduction in HbA1c levels and that evidence is insufficient to support the notion that CGMS provides a superior benefit over self-monitoring for HbA1c reduction. There was some indication from this review of improved detection of asymptomatic nocturnal hypoglycemia in the CGMS group. The 2007 Standards of Medical Care by the American Diabetes Association (ADA) does not mention this technology in the section on assessment of glycemic control. Recommendations in this section are for self-monitoring of blood glucose 3 or more times daily for patient using multiple insulin injections. The 2008 Standards of Care from the ADA include a recommendation that “CGMS may be a supplemental tool to SMBG for selected patients with type 1 diabetes, especially those with hypoglycemia unawareness.” This recommendation is level E, based on expert consensus or clinical experience. In December 2007 the Juvenile Diabetes Research Foundation (JDRF) completed recruitment for a 6-month trial at 10 centers of real-time CGMS in patients with type 1 diabetes. Results of this study, that randomly assigned 322 adults and children with type 1 diabetes to continuous glucose monitoring or self (home) monitoring, were released in 2008. With HbA1c as the primary outcome measure, there was a significant difference among patients 25 years of age or older that favored continuous monitoring (mean HbA1c difference 0.53%), while the difference between groups was not statistically significant for those age 15 to 24 years or 8 to 14 years. Unlike many prior studies, this study was sufficiently large to detect a meaningful change in HbA1c levels between groups. The population in this study had relatively well controlled diabetes in that entry criterion was glycated Hb of 7 to 10% but about 70% had levels between 7 and 8%; in addition, over 70% of patients were using an insulin pump. No significant differences were noted in rates of hypoglycemic events, but the study was likely not sufficiently large to detect potential differences. The authors also reported that monitor use was greatest in those patients age 25 or older where 83% of patients used the monitor 6 or more days per week. Summary In summary, the available studies demonstrate that intermittent glucose monitoring provides a different type of data than results from finger stick glucose levels. In addition to providing more data points, it also provides information about trends (direction) in glucose levels. This additional information is most likely to benefit those patients with type 1 diabetes who do not have adequate control, including episodes of hypoglycemia, despite use of current best practices including multiple (4 or more) daily checks of blood glucose and multiple (3 or more) insulin injections or use of an insulin pump. Thus, based on the available data and supported by clinical input, the policy statement is changed to indicate that intermittent, i.e., 72-hour, glucose monitoring may be considered medically necessary in those whose diabetes is poorly controlled despite use of best practices. The recently published study noted above, indicates that for those with type 1 diabetes mellitus over the age of 25 whose HbA1c is near (i.e., 7 to 8%), but not yet at target levels (generally less than 7%), use of continuous monitoring can result in further improvements in glucose control. However, it seems reasonable to extrapolate these results to adults (those over age 18) with HbA1c levels already near the target level. Data to support use of continuous glucose monitoring in other age groups or in other clinical situations are lacking. Using rationale similar to that noted above for intermittent monitoring, continuous monitoring can also be used, and can be considered medically necessary, to provide additional data for management of those who have recurrent unexplained symptomatic hypoglycemia, despite use of current best practices, that puts the patient or others at risk and for pregnant type 1 patients with diabetes. Data to support use (that show improved outcomes) of devices that allow wireless connectivity between a continuous monitoring device and insulin pump are still lacking. Procedure Codes and Billing Guidelines:
*Prior Approval is recommended for this policy. **Current Procedural Terminology © 2008 American Medical Association. All Rights Reserved. Wellmark medical policies address the complex issue of technology assessment of new and emerging treatments, devices, drugs, etc. They are developed to assist in administering plan benefits and constitute neither offers of coverage nor medical advice. Wellmark medical policies contain only a partial, general description of plan or program benefits and do not constitute a contract. Wellmark does not provide health care services and, therefore, cannot guarantee any results or outcomes. Participating providers are independent contractors in private practice and are neither employees nor agents of Wellmark or its affiliates. Treating providers are solely responsible for medical advice and treatment of members. Our medical policies may be updated and therefore are subject to change without notice. |
 |
Top