Response to Comments: Pneumatic Compression Devices - DL33829

One commenter wrote that “about 15% of patients with Critical Limb Ischemia (CLI) do not have options for revascularization. It is in this group that we have had success in preventing amputations, providing relief from ischemic rest pain and healing wounds by using compression therapy [...]. Utilizing fluorescence Angiography as well as Near Infrared Spectroscopy, we have been able to provide objective evidence that the compression does improve tissue perfusion and tissue oxygenation. The cost of this therapy is minimal compared to the cost of and amputation with its mortality morbidity and ongoing costs of rehabilitation.”

The following citations for supporting literature were submitted:

• Treatment of Non-Reconstructable Critical Limb Ischemia Utilizing Noninvasive Pneumatic Compression device Monitored with Fluorescent angiography. Wounds July 2018
• Ponticello M, Andersen CA, Marmolejo (Schade) VL. Limb Salvage Versus Amputation: A Closer Look at the Evidence, Costs and Long-Term Outcomes. Podiatry Today. March 2016; 29(3).
• Schade VL, Andersen CA, Omana-Daniels R, How Fluorescence Angiography Illuminates the Potential For Limb Salvage. Podiatry Today. September 2015; 28(9): 20-24

Thank you for the comments. However, two of these publications were review articles, and one was a case study with an extremely low number of patients (2). The publications reviewed imaging techniques and tissue perfusion, and did not review clinical outcomes in the Medicare-age population. We excluded such variables from our literature analysis due to their indirectness in predicting patient-centered outcomes in the Medicare-aged population.

The American Professional Wound Care Association proposed, “… that PCDs used to address CLI be considered a covered service.” They further stated that “... this is a tool APWCA members use to heal wounds, prevent amputations and decrease costly wound-related complications experienced by Medicare beneficiaries.”

Thank you for the comment. The DME MACs appreciate the advocacy of these commenters on behalf of Medicare beneficiaries. However, no additional clinical literature was provided to support Medicare requirements for coverage.

One commenter wrote about their practice experience, noting “We currently have eight patients who have been on the ArtAssist pump for at least three months. Of these, all have shown improvements in their VascuQol scores. Many of the initial follow-up score changes were dramatic, with several initial scores in the 1-2 range improving to 4-6 range (scores range from 1-7, with 7 being the best score). None have had a major amputation.”

They added "... we did attempt ArtAssist for seven other patients, but they were only on the pumps for a limited time (1-4 weeks) prior to either going on to amputation or expiring. We believe that they presented too late in their disease course for the ArtAssist to have time to develop collateral circulation. Of note, several of these patients did experience a decrease in rest pain prior to expiring, which at least improved their quality of life.” The author also stated that “patients deserve an alternative to amputation when no endovascular or surgical options exist. The ArtAssist pump is a viable option, but not available to most patients due to cost. Providing coverage for this relatively low-cost device would allow for those patients to at least have a chance at limb salvage.”

Thank you for the comments. However, no additional clinical literature was provided to support Medicare requirements for coverage. Moreover, the small numbers cited by the commenter limit policy-wide applicability.

The Alliance of Wound Care Stakeholders noted “We urge that the DMEMACs again provide coverage for E0675 for the treatment of patients with CLI for both the upper and lower extremities. We are in agreement with the points outlined in the requestor’s comment letter that counter the reasons that the DME MACs decided against coverage of these devices. We would appreciate that you review carefully their points regarding but not limited to sample size, high probability of bias without sham controls and blinding, generalizability to the Medicare population, exclusion of co-morbidities.”

Thank you for the comments. However, no additional clinical literature was provided to support Medicare requirements for coverage.

One beneficiary commented that “I have been using the ArtAssist pump for eight months and my doctors are impressed with how my foot and toes look […]. I was disappointed that Medicare would not cover it, but I bought it anyway with hope it would save my other foot and hope Medicare will embrace the ArtAssist pump as it has worked for me.”

The DME MACs appreciate your comments regarding your personal experiences with this device.

Several clinicians, one on behalf of a specialty society, submitted comments advocating coverage of PCD and pointed to Reference 5 in the bibliography of the proposed policy, which stated that “Results from multiple randomized controlled trials suggest that limb compression is beneficial in improving ACD. Further, numerous supplied studies have shown that arterial flow, cutaneous blood flow, growth factors and NO are all significantly enhanced. Human trials do confirm the above hypothesis.”

According to the commenter, the draft LCD discounts physiological and anatomical studies and refuses to use them to evaluate efficacy, even in situations wherein efficacy is proven, the LCD then dismisses the studies for having a small sample size. “However, when reviewing the evidence provided to rebut this notion, one can see that the average sample size for the human clinical trials was at least 34 participants with a p < 0.05. The LCD harbors the narrative that even if the appropriate sample size were selected for a clinical trial, there would still be an inappropriate assessment of the target population. Why? The LCD stipulates that patients studied do not represent Medicare eligibility. The most susceptible patient populations are unlikely to improve outcomes without first improving access to higher intensity quality care measures. However, as one may dig deeper into the target population, mostly African Americans; the clear concise is that these communities are far less likely to access more advanced surgical options for limb salvage. Thus, given a proper workup for CLI, providing patients with all available treatments to prevent amputation makes good sense. Finally, the DME MACs have indicated that they respond to a collection of vascular and surgical specialties with extensive clinical and academic research experience. However, by failing to accept the recommendations made by key stakeholders such as SVM [Society of Vascular Medicine] and SIR [Society for Interventional Radiology], then DME MAC medical directors are reinforcing the lower quality of life for patients and leaving them with only one option, amputation.”

Thank you for the comments. However, the DME MACs respectfully disagree. The studies we reviewed and analyzed outlined in our bibliography regarding Intermittent Pneumatic Compression (IPC) typically excluded patients with co-morbidities and conditions common within the Medicare patient population.

The review article cited by the requestor (Reference 5 in the bibliography) noted “We restricted our review to studies that evaluated patients with intermittent claudication and therefore, did not include studies that looked at the effect of compression therapy on physiologic measures in asymptomatic patients or patients with critical limb ischemia.” [emphasis added] Thus, we are unclear on part of the comment regarding critical limb ischemia.

The DME MACs note that Oresanya et. al. (Reference 5 in bibliography) concluded that “Results from multiple randomized controlled trials suggest that limb compression is beneficial in improving ACD.” However, the authors also acknowledged that studies included in their systematic analysis, “had multiple exclusion criteria, such as excluding patients with comorbid conditions or expected low compliance, limiting the generalizability of the results.” The authors also wrote that the “overall quality of included studies were poor.” The authors further tempered their conclusion of benefit stating that “there is a need for further studies…to better define the role of PCD in the treatment of claudication.”

The non-coverage determination was made due to a lack of certainty regarding PCD benefit as well as a lack of generalizability to the Medicare population as summarized in the policy.

Further comments regarding the studies in the published peer-reviewed literature are enumerated in our response to Comment 9 below.

One commenter wrote that “The draft LCD stated that ‘… most qualified their conclusion stating a need for additional studies.’ These statements do not dilute previous conclusions of efficacy but usually call for additional studies to provide further insight into the field of study so that further advancements can be made.”

Thank you for the comment. Please see our response to Comment 6 above, and Comment 9 below.

One beneficiary advocating for coverage of PCD wrote: “I am writing on behalf of Medicare’s request for people who have used the Art Assist to let you know my experience with the machine.

About three summers ago I noticed that my index finger was turning a deep shade of bluish black and the nail was degrading. I have had Raynaud’s in my hands and feet for years due to Lupus, but never anything like this. After visiting a doctor and having a test to measure the blood flow in my arm and hand, it was discovered that an artery was [occluded] (closed off) and starting to affect my ring finger as well. I was referred to Mayo Clinic in Rochester, Minnesota and they confirmed that I also have Scleroderma which led to the artery issue. They recommended the Art Assist and to come and use it as much as possible. Since I live in Montana, that was not possible, so I be researching the product and was able to get one sent to me by making payments. I was told Medicare would not pay for it.

I have been using the Art Assist Machine for over two years now and you can’t tell the difference between the originally affected finger and my normal ones. My fingers all look healthy and there is no discoloration except when I forget it’s winter and don’t remember to wear my gloves. My fault, not the machine! If I get busy and skip a couple of days, my finger hurts-pulses, and reminds me to get on the machine! It is a ‘Use or [lose] situation’.

I am blessed to have been able to get this machine and cutting corners was worth it in order to pay for it. But not everyone can do that. I highly recommend this machine for hands or feet and that Medicare strongly consider covering some of the cost for people whose only other option is pain and amputation.”

The DME MACs appreciate your comments regarding your personal experiences with this device.

One commenter on behalf of a specialty society stated that “The draft LCD states that in the 7th paragraph on page 8: IPC has been previously shown to increase the arterial flow rate in the popliteal artery and is thought to promote the release of angiogenic growth factors and endothelial nitric oxide, an intrinsic vasodilator. However, human clinical studies do not confirm these hypotheses (5) Reference 5, reaches an opposite conclusion stating, ‘Results from multiple randomized controlled trials suggest that limb compression is beneficial in improving ACD.’ Further, numerous supplied studies have shown that arterial flow, cutaneous blood flow, growth factors and NO are all significantly enhanced. Human trials do confirm the above hypothesis.” Supporting literature was not submitted.

The DME MACs appreciate your comments and advocacy of specialty societies on behalf of Medicare beneficiaries. However, while the meta-analysis was performed with statistical integrity based on the sensitivity analyses performed, the absence of sham arms and pooling of heterogeneous studies of poor quality and small sample sizes significantly limits the certainty of the IPC treatment-effect size. Moreover, the certainty of this meta-analysis is even further limited by a high risk of bias due to the absence of sham arms and unblinded participants notable of the individual studies included. Furthermore, the mean improvement in ACD of 125 m is well below the minimally important difference of 305 m for improvement identified in CLI patients. Finally, as noted in the response to Comment 6, the authors of this publication noted limited generalizability of the results due to multiple and significant concerns.

One commenter noted that “The draft LCD states that studies were of small sample size (<10). The meta-analysis included eight RCTs, each with experimental sizes of 20 or more. There were 34 patients in the Mayo Clinic limb salvage study (3), 62 patients with 73% amputation free survival at 3 years (34), 171 patients with 94% limb salvage at 3.5 years (59), Twenty limbs with CLI studied with significant increases in arterial and skin blood flow [reference submitted]. Studies are considered small, or efficacy is not obtained when P values are above 0.05. All arterial IPC studies have shown P values that support the experimental (pump) group. P values account for sample size in its calculation so the statistical significance is met when P<0.05. When an experimental group demonstrates a low P value with a small sample size, this means that the effect of the pump is large compared to the control group. The effect of arterial IPC is large.”

Thank you for the comments. However, the DME MACs respectfully disagree. We have responded to similar comments regarding the review article(s) and quality of studies above (see Comment 6 and Comment 9).

The 2008 Mayo CLI study was excluded from the review article (Reference 5 in the bibliography), for reasons noted in response to Comment 6. This was a small retrospective study that had 27 patients in the intervention arm, and concluded that “This controlled study adds to the momentum of IPC clinical efficacy in critical limb ischemia set by previously published case series, compelling the pursuit of large scale multicentric level 1 studies to substantiate its actual clinical role, relative indications, and to enhance our insight into the pertinent physiologic mechanisms.” [emphasis added] No such subsequent study has been carried out.

The draft LCD concluded that most of the studies reviewed were of small sample size. Small sample sizes are statistically more likely to yield results that are not replicated in subsequent larger trials, and thus, those results may not represent results that would be obtained from the much larger intended population – in this case – Medicare beneficiaries. Furthermore, the level of confidence in the results and conclusions obtained from small studies is limited.

One commenter wrote that “The draft LCD states that there is high probability for bias without sham controls and blinding:

a. In 2004, before publications of the Mayo studies [references submitted], Sultan studies [references submitted], an IRB approved, sham controlled trial was conducted but never published due to a non-study related issue. Data from this trial were presented at a vascular surgical conference. Those results combined with the many later published studies have prompted recent ethical concerns from the primary submitter’s IRB that there is likely lack of equipoise for a sham controlled critical limb ischemia, no-option limb salvage trial. b. While prospective randomized sham-controlled studies provide valuable data when effectively executed we believe that the above referenced studies provide compelling evidence in support of IPC in critical limb ischemia. Since placing an IPC pump on a leg and having it cycle, whether at full strength or sham, is obvious to the patient. Creating a sham would require some degree of compression and we have no data on how low of a compression strength would be suitable. Also, although IPC is low risk some IRBs have concern about using a sham device which could potentially cause a mechanical injury and would require that the patient be seated and “treated” for 2-3 hours daily. For the sham group this inconvenience is not met with any potential benefit and thus suboptimal. c. None of the study centers were financially supported by manufacturers of pump devices which reduced possible financial incentive bias. d. Measurements used in the submitted studies (limb salvage, TcPO2, laser Doppler skin flow, duplex ultrasound automatically calculated values, treadmill times and distance are quantifiable, objective, not easily altered for preferential outcomes. e. Sham controlled studies were not used in evaluating currently accepted as reasonable and necessary therapeutic modalities including supervised exercise for intermittent claudication, angioplasty devices and associated procedures.” 

Thank you for the comment. However, the DME MACs respectfully disagree that there is compelling evidence in support of IPC in critical limb ischemia.

Our response to Comment 10 above, discusses the Mayo retrospective study cited in this comment. The 2008 Sultan study was a very small non-randomized trial without any controls. The 2011 Sultan publication was a single center non-randomized trial without controls, which was performed in another country whose population is quite different from our Medicare population.

Clinical studies seek to demonstrate the effect attributable to an intervention. Sham procedures are important, especially in interventional studies, where they more clearly help distinguish whether an intervention is effective beyond the placebo response. For example, see Reference 69 in the final policy, regarding sham placebo effects on objectively measured clinical responses. We agree that a sham intervention would require some degree of compression and that there is no data on how low of a compression strength would be suitable. However, the argument does not negate our concerns that there is a probability of bias/placebo effects, and that there is insufficient evidence regarding the effects attributable to PCD.

One commenter wrote regarding a concern that “The draft LCD is discounting physiological and anatomical studies including physiological studies of patency and blood flow, arterial evidence of collateral growth are primary measurements. These studies should be used in evaluation of efficacy. a. Physiological studies of patency and blood flow are primary outcome measurements in evaluation of endovascular devices and procedures such as angioplasty balloons, stents and atherectomy devices. b. Two submitted studies show arteriographic evidence of collateral growth (studies were referenced), a long-term result that supports the durable results seen in both CLI and claudication studies. Collateralization is the reason for durability of the IPC treatment seen in CLI studies [references submitted] and in the claudication trials (studies were referenced). c. Submitted studies measured significant increases in biochemical factors in support of collateral arterial growth, acute blood flow increases at several arterial levels, foot skin flow where most ischemic ulcers present and in providing beneficial hematologic effects such as increased nitrites (NO), increased tissue factor pathway inhibitor, decreased PAI-1 (plasminogen activator inhibitor). d. Numerous studies have identified shear stress as a regulator of endothelial cell function and vascular health. The rapidly applied pressure rise of arterial pumps, as compared to pumps used for DVT prophylaxis and venous/lymphatic treatment, creates endothelial shear stress that is largely diminished in obstructed arteries. e. Discounting these important studies detracts significantly from the total evidence picture that supports use of this modality as we recommend for no-option CLI patients.”

Thank you for the comments. However, the DME MACs respectfully disagree. These variables were excluded from our literature analysis due to their indirectness in predicting patient-centered clinical outcomes.

One commenter stated that “Several submitted studies show durability of outcome up to 3.5 years. This is at least as good as durability expected with endovascular procedures.”

Thank you for the comment. However, the DME MACs respectfully disagree. While these studies included patients of Medicare-age, the patient-exclusion criteria included co-morbidities that are typically observed in the Medicare population, therefore, the outcomes observed may not be applicable. Moreover, we have also discussed the limitations of the submitted studies when responding to several comments above.

One commenter wrote regarding data insufficiency stating that “The submitting physicians with support from several pertinent specialty societies are asserting that, contrary to the opinion expressed in the LCD draft, there is adequate evidence of sufficient data that arterial IPC is safe, effective, reasonable and necessary to provide our patients with a viable option to amputation where no other options are available.”

Thank you for the comment. However, we respectfully disagree. In order to meet Medicare coverage guidelines, there must be peer-reviewed clinical literature which demonstrates the safety and efficacy of the treatment in the Medicare population. As we have pointed out in other responses above, the requestor and other commenters did not submit any randomized controlled trials or other good quality adequately powered trials that support the beneficial effects of IPC in patients with critical limb ischemia.

One commenter wrote that “The concern that patients studied do not represent those of Medicare eligibility is misplaced to the extent that almost all patients studied with CLI were of Medicare age 65 or over:

Kavros (3) 68.7- 71.3 IQR; Labropoulos (45) range of 65 to 83 years; Sultan (59) 68 to 81 years IQR; Montori (52) 63-79 IQR; van Bemmelen (65) 76.2 mean age.”

Thank you for the comment. However, the DME MACs respectfully disagree. The studies in our bibliography include the articles you cited in your comments. The patient-exclusion criteria in these studies included co-morbidities that are typical for the Medicare population, and therefore, the outcomes observed may not be applicable.

Another commenter wrote that a “... statement in the LCD draft that relates to African American communities that have less access to more advanced surgical options for limb salvage. Contrary to the LCD draft’s expressed opinion that improvement in outcome is ‘unlikely’ without access to surgery it is certainly reasonable and necessary to provide this option where access to existing treatments is less available. Given a proper workup for CLI, it makes good sense to provide patients with all available treatments to prevent amputation which we anticipate will increase access to care for disadvantaged groups such as women and racial minorities.”

Thank you for the comments. However, no clinical literature was provided to support Medicare requirements for coverage.

One commenter noted that “There is no data that rural, Hispanic, African American or any other identifiable population have different anatomy, physiology, or biochemistry, regarding PAD, that would suggest any different result from the populations studied.”

Thank you for the comments, and we agree in-part with your comment regarding lack of data on differences in anatomy, physiology, and biochemistry. However, sometimes other more important variables, such as health care disparities and social determinants of health, are the primary drivers of clinically meaningful, and patient centered health outcomes in the Medicare population.

One commenter stated that “All tested devices apply close to systolic pressures to patients in the seated position, reaching pressures much more rapidly than existing IPC types, for durations of 3 to 4 seconds and with three cycles per minute. This is far more homogeneous than the existing IPC types for venous and lymphatic treatment.”

Thank you for the comments. However, no clinical literature was provided to support Medicare requirements for coverage. Moreover, this reconsideration exercise is not a comparative one with regards to other conditions and treatment modalities.

One commenter wrote that “The specialty society recommendations in the draft LCD are dated 2016 and 2019. More recent society recommendations, with benefit of member input now endorse use of arterial IPC in no-option CLI patients. The Society of Vascular Medicine now supports the LCD request with unanimous consent of its board. SVM is best situated to evaluate this medical modality and is without possible conflicts of interest from surgical organizations. Other specialty societies supporting the call for arterial pump coverage that represent physicians who work with non-reconstructable, no-option CLI patients include the Society of Interventional Radiology, The American Board of Wound Medicine and Surgery, the American Professional Wound Care Association, The Save A Leg Save A Life Foundation, the Association for the Advancement of Wound Care and the Alliance of Wound Care Stakeholders (submitted separately). Letters of specialty society recommendations are attached.”

The DME MACs appreciate your comments and the advocacy of specialty societies on behalf of Medicare beneficiaries.

The society guidelines cited in the proposed policy are the most recent ones available in the peer-reviewed literature, and the letters submitted in support of this comment did not cite any more recent clinical guidelines.

Although the society guidelines cited in the draft LCD do discuss the use of arterial IPC in no-option CLI patients, no target population was identified, no clear clinical outcomes were identified, and the level of recommendation(s) was weak (IIB) indicating uncertainty.

As noted above, no additional or more recent peer-reviewed clinical guidelines, other than the ones cited in the draft policy were submitted to support coverage in the Medicare-age population.

One commenter noted that “The submitted IPC studies have reasonably excluded many of those Medicare eligible patients that would not be candidates for this therapy.”

Thank you for the comment. We concur. The DME MACs identified studies which have included Medicare eligible patients. However, the typical Medicare age beneficiary is usually aged 65 or over, and with multiple co-morbidities. The submitted studies excluded the "typical" Medicare patient, hence our conclusion that the available clinical literature does not establish the efficacy of this treatment in the general/typical Medicare-age population.

One commenter stated that “The Veterans Administration’s Preservation-Amputation-for-Veterans Everywhere (PAVE) Program was created in 1993 to prevent limb amputations by providing the best possible care to ensure the Veteran receives optimal and compassionate patient-centered care by implementing evidence-based prevention practices (VHA Directive 1410). Within that framework, The VA medical centers have been using arterial IPC devices for the past 22 years for approximately 6000 patients according to one of the several suppliers of devices. This demonstrates that the often-overwhelmed VA system sees efficacy and value in providing this treatment modality.”

Thank you for the comment. However, the VA and Medicare are two distinct programs with different rules, regulations and standards for coverage of items and services. Moreover, the VA population of patients does not necessarily reflect that of Medicare-eligible beneficiaries.

One commenting group said that they “... appreciated the statement that ‘… the DME MACs are unable to make recommendations for or against IPC with any certainty.’ However, our group of responders is a collection of vascular medical and surgical specialists with extensive clinical and academic research experience. The published research alone is sufficient for those of us that treat these patients to have established efficacy for this patient population. Additionally, our experience with this modality greatly reinforces our comfort in strongly recognizing its need for these unfortunate patients with no likely option other than amputation.”

Thank you for the comments. The clinical literature cited in the proposed LCD and/or submitted by commenters above does not validate the certainty of efficacy in the Medicare population.

No additional clinical literature was submitted by this commenting group to support Medicare requirements for coverage.