In April 2009, a 4-day conference titled “ESRD: State of the Art and Charting the Challenges for the Future” was held in Boston (conference report and steering committee letter). The main purpose of the conference was to address the lack of improvement in dialysis patient survival over the last twenty years. At that meeting, sudden cardiac death was identified as the #1 cause of death for dialysis patients, accounting for 59% of cardiovascular-related deaths.
At the end of the conference, a novel finding was reached. It was concluded that these cardiovascular-related deaths were not due primarily to atherosclerotic (plaques and arterial stiffening) disease, but rather uremic cardiomyopathy, characterized by left ventricular hypertrophy (LVH), LV dysfunction, and LV dilatation. The enlarged muscle walls of the left ventricle become fibrotic and fail to conduct electrical impulses correctly due to repeated and continual fluid overloads in the body (most dialysis patients don’t produce urine). The conference’s conclusions have driven many efforts involving several leading nephrologists to reduce sodium levels in dialysate, end sodium profiling, improve body water volume control, and more.
Among the many members of the steering committee of the conference was Raymond Hakim, MD, PhD, then Chief Medical Officer of Fresenius Medical Care North America (FMC). The Boston conference’s findings regarding sudden cardiac death appear to have driven investigations within FMC. This is very evident in the 6-page internal memo from Fresenius Medical Services (FMS) to FMC’s medical directors and attending physicians dated November 4, 2011. The memo’s subject line is “RE: Dialysate Bicarbonate, Alkalosis, and Patient Safety.” While the memo does not bear his name, Dr. Hakim was head of the FMS at that time. This memo refers to a FMCNA analysis, a case-control study that evaluated risk factors in hemodialysis patients who suffered from cardiopulmonary arrest in the dialysis facility compared to other hemodialysis patients within the same facilities between Jan. 1 and Dec. 31, 2010.
This case-control study identified 941 patients in 667 FMC facilities between Jan. 1 and Dec. 31, 2010 that had cardiopulmonary (CP) arrests within the facilities. Looking at the data for these 941 patients, the study found a 4.7 times higher risk (adjusted) of CP arrest with pre-dialysis bicarbonate levels >28mEq/L. If the patient also had a pre-dialysis potassium that was <4 mEq/L, they had a 6.3 times higher (adjusted) risk of CP arrest.
Before 2011, alkalosis, routinely measured via serum bicarbonate levels, had not been seen a significant problem for hemodialysis patients. An earlier non-FMC study (2001-2003, DaVita data) showed some survival benefit with pre-dialysis bicarbonate levels >24 mEq/L.
Working Hypothesis and/or Preliminary Concern Stage
The internal FMS memo released Nov. 4 contains several references to FMS efforts and programs to educate its medical staff on this issue starting in January 2011. Before the results of the case-control study were published by FMS on Nov. 4, 2011, there were likely preliminary data analyses, as well as internal communications and educational programs done as part of a preliminary concern or working hypothesis stage.
The Nov.4 memo mentions that several memos had previously been sent to FMC staff explaining the differences between Naturalyte® and Granuflo®, both FMC acid concentrate products. It also states that recommendations had previously been made to address pre-dialysis alkalosis found in an increasing number of patients. The Nov. 4 memo specifically states that this problem was mentioned in the Medical Staff Newsletter in January 2011 and that two recent presentations by FMC personnel were available for download from the “Doctors Corner.”
It appears that the FMS division of FMC concluded in October or early November 2011 that their internal data supported a conclusion that a significant risk of cardiopulmonary arrest existed for patients with the use of Granuflo. It was no longer simply a working hypothesis. On Nov. 4, 2011, the FMS office released the 6-page memo, entitled “Dialysate Bicarbonate, Alkalosis, and Patient Safety,” that emphatically stated the need to take action if a patient’s pre-dialysis serum bicarbonate was >28 mEq/L and especially if the patient also had a pre-dialysis serum potassium of <4 mEq/L. It also recommended “immediate attention to decreasing the prescribed dialysate bicarbonate in patients with pre-dialysis bicarbonate level of >24 mEq/L.
Statements and Warnings Regarding Alkalosis in the Internal FMC Memo
The memo begins with a one paragraph conclusion:
"Recent analyses performed using FMCNA hemodialysis (HD) patient safety data confirms that alkalosis is a significant risk factor associated with cardiopulmonary (CP) arrest in the dialysis unit, independent of and additive to the risk of CP arrest associated with pre-dialysis hypokalemia. The major cause of metabolic alkalosis in dialysis patients is inappropriately high dialysate total buffer concentration. As recommended in previous communications, physicians should individualized dialysate bicarbonate and total buffer prescriptions. We further recommend that predialysis serum bicarbonate level of >24 mEq/L should prompt immediate review of dialysate bicarbonate prescription."
The recommendations section of the memo begins by stating that “pre-dialysis alkalosis and hypokalemia are modifiable risk factors associated with CP arrest.” The memo urges that this issue “needs to be addressed urgently.”“The current analysis determined that: “borderline elevated pre-dialysis bicarbonate levels and over alkalosis are significantly associated with 6 to 8 fold greater increase of cardiopulmonary arrest and sudden cardiac death in the dialysis facility.” “ (italic type emphasis is theirs)
“In light of these troubling findings, we strongly recommend that physicians adjust dialysate bicarbonate prescriptions monthly for individual patients, with immediate attention to patients with serum pre-dialysis bicarbonate level of >24 mEq/L.”
The memo acknowledges that many facilities have converted to Fresenius powdered Granuflo formulations and that the total buffer level equals “prescribed bicarbonate plus 8.” It continues:
Among the ten cited references at the end of the memo are two medical articles that include “sudden death” in their titles. Another is entitled, “The faster potassium-lowering effect of high dialysate bicarbonate concentrates in chronic hemodialysis patients.”“There are instances whereby the physicians’ bicarbonate prescriptions were kept the same when shifting to power (sic) concentrate (Granuflo) (failing to account for the additional 8 mEq/L of sodium acetate), thus exposing patients to a higher total buffer load than intended. While 60% of current dialysate prescriptions are for 37 mEq/L of bicarbonate, it may be prudent to initially target a prescription of 31-33 mEq/L of dialysate bicarbonate (with total buffer greater by up to ~8 mEq/L from acetate) and adjust accordingly to patients’ monthly bicarbonate level. Please recall also that an additional source of bicarbonate may be the phosphate binders that are prescribed to patients.”
Failure to Notify Either the FDA or Non-FMC Dialysis Centers Using Granuflo
As noted earlier, Granuflo with sodium diacetate had been approved for marketing by the FDA in 2003. The problem with Granuflo, alkalosis, and sudden cardiac death was apparently not suspected until many years later, apparently around 2010.
Currently there are over 5700 dialysis centers and approximately 400,000 dialysis patients in the United States. It is estimated that 3300 of these clinics use Granuflo formulations and that approximately 260,000 patients are using the product. It is estimated that slightly less than half these patients, or approximately 125,000 patients, are using Granuflo in non-FMC clinics. Total revenues from Granuflo sales per year in the U.S. are estimated at a minimum of $80 million.
If a company fails to notify the FDA of a known problem with an FDA-approved medical device, it can result in severe fines and penalties, including being prohibited from marketing products for use in facilities that treat Medicare patients.
It is believed that the FDA received information concerning the alkalosis problem with Granuflo, most likely the internal FMC memo dated Nov. 4, in early 2012 from an anonymous source. The FDA then contacted all four dialysis concentrate manufacturers, FMC, Rockwell Medical, Diasol and Minntech Renal Systems, for information on this topic on March 27, 2012.
There is no evidence that FMC attempted to contact either the FDA or any of the non-FMC clinics until March 29, 2012. On that date, the FMC medical products division provided a 2-page memo to non-FMC clinics that used much of the same language from the FMS division memo of Nov. 4, 2011. The memo from the medical products division, however, lacked most of the detail and all of the urgency of the FMC internal memo of Nov. 4, 2011 on the same topic.
- The March 29 memo only included four medical resources on its reference list.
- The March 29 memo included only one of the ten references used in the FMC internal memo of Nov. 4.
- One of the references on the March 29 memo only was for 2001-2003 data that showed a benefit of pre-dialysis serum bicarbonate levels of >24 mEq/L. (CJASN 2006 1:70-78)
- No references in the March 29 memo mentioned sudden death or the rapid removal of potassium during dialysis treatments with high bicarbonate levels in their titles.
Dr. Raymond Hakim Steps Down as FMC Medical Director
As previously mentioned, Dr. Raymond Hakim was the medical director of FMS on Nov. 4, 2011 and was likely responsible for the FMS internal memo released on that date. As Chief Medical Officer, it would seem logical that Dr. Hakim would also notify his non-physician superiors at FMC, Ben Lipps and (Mau)Rice Powell, of his decision to release the Nov 4, 2011 internal FMC memo. He would likely also have recommended that the medical products division of FMC notify their other customers, the non-FMC clinics, of the FMS findings so these non-FMC clinics could make the best decisions for their patients regarding the use of Granuflo.
Neither Lipps nor Powell are physicians (current FMC management board members).
Six days after the internal memo was released, the annual meeting of the American Society of Nephrology (ASN) was held in Philadelphia (Nov. 10-12). At that meeting, there was a Dialysis Outcomes Practice Patterns Study (DOPPS) presentation (click here) on bicarbonate levels and dialysis patient mortality. It concluded that there was an increase in patient mortality with total buffer levels above 37 mEq/L.
It appears that Dr. Hakim left Philadelphia more convinced than ever that non-FMC clinics needed to be notified of his finding concerning Granuflo, alkalosis, and CP arrests. It would seem logical that Dr. Hakim would become more adamant in his requests that the product division of FMC do this as soon as possible.
On Nov. 16, 2011, a press release appeared from FMCNA stating that Dr. Raymond Hakim was stepping down as FMC Medical Director as of December 12, 2011. It was also announced that he would be replaced by Dr. Frank Maddux.
Did FMC Put Market Share and Profit before Patient Safety?
It appears that Fresenius Medical Care (FMC) knew as early as 2010 that there may be a link between high total buffer levels and the risk of cardiac arrest. In its response to the Part I posting on RenalWEB about this issue, FMC said it didn’t believe it had enough information to share the potential risks of Granuflo with its customers:
But is there not a dangerous conflict of interest for a company that treats patients and makes the products as well? When clinical experts bump heads with those who want to protect market share and protect profit margins, who makes the final decision?""In this case, with respect to 'total buffer,' the Chief Medical Office distributed to FMS medical directors a working hypothesis relating to dialysate mix products in general, but also specifically citing our products Granuflo and Naturalyte, that -- pending further data collection, analysis, and development -- might or might not eventually support conclusions warranting general external or peer reviewed publication."
FMC started as a products company; its top two executives, who are not physicians but instead medical product sales and development experts, may have, in this case, agreed to address these medical concerns with their own patients, but hesitated about informing their customers, possibly fearing loss of market share and possible litigation. As a result, with the current FMC medical product oversight policies, it appears that non-physician FMC executives can now make important medical decisions concerning dialysis patients in non-FMC clinics.
Clearly, the two-page memo sent to customers was devoid of much of the information in the internal memo. It was purposely made vague, when the evidence was crystal clear: high pre-dialysis serum bicarbonate levels created a risk of cardiac arrest and deaths. FMC wanted to limit its risk and, at the same time, not scare off customers.
This is a story not about just one physician's conflict of interest, but about multiple persons' conflicts of interest in a vertically-integrated, for-profit healthcare corporation. All these conflicts of interest had a multiplying effect, resulting in risks and dangers for patients that are at a minimum, highly unacceptable, and at worst, criminal.
Many of the hypotheses in this article cannot be confirmed due to corporate confidentiality policies and the non-disclosure clauses of employment and/or employment termination contracts. Only congressional hearings or an investigation by the HHS Office of the Inspector General can bring these facts to light.
A group of dialysis clinic medical directors of a large dialysis provider corporation with approximately 13,000 patients held a regularly scheduled meeting on Friday, May 11, 2012. One of topics for discussion was the 2-page, March 29 memo from FMC. One physician who had reviewed the document said that the memo's "conclusion was not supported by the memo's references." The group decided to take no patient-specific action, but did decide to choose to draw post-dialysis serum bicarbonate levels on some patients, with the results to be discussed at a later time.
Unrelated to this, I received copies of both the internal and external FMC memos on Saturday, May 12. Within a hour of receiving them, I contacted the CEO of this large dialysis provider. He then had one of his physicians who had privileges at an FMC clinic obtain a copy of the internal FMC 6-page memo dated Nov. 4.
After reading the 6-page memo, these medical directors decided to meet again on Monday, May 14. At that meeting, they decided to intervene in approximately 20% of their patients care. On Tuesday, May 15, patient-specific changes were implemented for over 2000 patients.