Skip to main content Back to Top

4/21/1998

Hemolysis and Renal Failure Associated with Use of Sterile Water for Injection to Dilute 25% Human Albumin Solution

Cheryl A. Thompson

FDA is aware of four cases of hemolysis, reported starting in 1994, that occurred during or after plasmapheresis with 25% human albumin solution diluted with sterile water for injection to a final protein concentration of 5%. Hemolysis was followed by acute renal failure in two of the cases. 

The hemolysis was attributed to infusion, during plasmapheresis, of a hypotonic solution of 25% albumin solution, inappropriately diluted with sterile water for injection. Albumin solutions of all concentrations are formulated to have a sodium concentration of 130–160 meq/L (1). Thus, a fivefold dilution of albumin with sterile water for injection produces a solution with osmolarity (or tonicity) of only about one fifth that of 0.9% sodium chloride injection (isotonic saline). Hemolysis has occurred from use of albumin solutions containing less than 90 meq/L of sodium, which would be the case when albumin 25% solution is diluted fivefold with sterile water for injection (2,3).

The large volumes of albumin solution used in plasmapheresis or plasma exchange were likely a contributing factor in the hemolysis cases reported to FDA. The hypotonic dilutions of albumin, once infused, would have accounted for a significant fraction of the patients’ calculated blood volumes. In a hypotonic environment, the shearing strain produced by the plasmapheresis devices may also have contributed to hemolysis. 

One of the four cases has already been published (4), and another was published just weeks ago (5). Additional clinical details of these cases—and details of two more cases reported to FDA’s MedWatch program—are summarized here. 

Case 1 involved a 64-year-old man undergoing plasmapheresis for treatment of myasthenia gravis. The procedure was started with 5% albumin solution that had been prepared by diluting one volume of 25% solution with four volumes of sterile water for injection; magnesium, potassium, and calcium salts had been added. After 1350 mL was administered over 50 minutes, the hemolysis alarm on the plasmapheresis device sounded. The procedure was interrupted for 20 minutes while new plasmapheresis lines were placed and primed and was then resumed. After 189 mL more of the diluted albumin solution was administered over 27 minutes, the alarm again sounded and the procedure was stopped. 

After plasmapheresis, the man had jaw and arm pain and hematuria. Renal failure ensued, requiring dialysis. His serum creatinine (SCr) concentration increased from a baseline of 0.9 mg/dL 2 days before plasmapheresis to a peak of 14.1 mg/dL 11 days after the procedure. The man subsequently recovered without additional treatment and without sequelae. Other patients who later underwent plasmapheresis with the same device—but with 25% albumin solution diluted with 5% dextrose injection or 0.9% sodium chloride injection to a final albumin concentration of 5%—had no problem. 

Case 2 is the index case referred to by Steinmuller (5). The patient was a 66-year-old man with a clinical course complicated by a history of multiple myeloma, hypercalcemia, hyperglobulinemia, d iabetes mellitus, hypoalbuminemia, and renal insufficiency. He underwent plasmapheresis, dialysis, and chemotherapy with vincristine sulfate and doxorubicin hydrochloride during his first few days in the hospital. The first plasmapheresis—with commercially prepared 5% albumin solution—was without incident. In the next four days, the man underwent two more courses of plasmapheresis but with 5% albumin solution prepared by diluting one volume of 25% solution with four volumes of sterile water for injection; magnesium, potassium, and calcium salts had been added. No alarm sounded on the plasmapheresis device during either course, and the patient had no signs or symptoms of hemolysis. However, the results of laboratory tests taken before the second plasmapheresis and immediately after the third showed the following changes: The hematocrit decreased from 30.0% to 22.7% and from 27.4% to 22.7% (packed red blood cells were transfused between the two plasmapheresis sessions), total serum bilirubin concentration increased from 0.4 to 8.8 mg/dL, serum lactate dehydrogenase concentration increased from 928 to 3116 U/L, and blood urea nitrogen (BUN) increased from 29 to 98 mg/dL and SCr concentration increased from 3.1 to 4.9 mg/dL (even with dialysis between plasmapheresis sessions). 

The man did not respond to chemotherapy or additional dialysis and died about two weeks later. Although the primary cause of death was attributed to complications of the man’s underlying conditions, hemolysis and deterioration in renal function after the second and third courses of plasmapheresis were considered contributory. 

Case 3 was described in the literature by Forte et al. (4). A 68-year-old man with immunoglobulin M gammopathy underwent plasmapheresis with 5% albumin solution. The albumin solution was prepared by dilution of one volume of 25% solution with four volumes of sterile water for injection. Hemolysis ensued, evidenced by red-tinged fluid in the plasmapheresis device’s tubing and bags. There was no other sign or symptom of hemolysis. No alarm on the plasmapheresis device sounded during the procedure. Plasmapheresis was immediately halted, and the patient was hydrated with 0.9% sodium chloride injection. He was admitted to the hospital for monitoring of possible sequelae and discharged within 48 hours. There was no appreciable change in the hematocrit or concentrations of BUN or SCr, and there were no other sequelae. 

Case 4 involved a 74-year-old man undergoing plasmapheresis for treatment of myasthenia gravis. The procedure started with 5% albumin solution prepared by diluting one volume of 25% solution with four volumes of sterile water for injection. Hemolysis occurred, evidenced by red-tinged plasma in the collection bag, red-tinged urine, dyspnea, and a decrease in oxygen saturation from 98% to 67%. Plasmapheresis was stopped and then resumed with commercially prepared 5% albumin solution without incident. By evening, the man’s total serum bilirubin concentration had increased to 5.4 mg/dL, from a preprocedure level of 0.4 mg/dL. His BUN progressively increased from a preprocedure level of 19 mg/dL to a peak of 43 mg/dL three days after plasmapheresis. The man recovered without sequelae. 

In all four cases, the hospital pharmacies relied on information contained in the seventh (published in 1992) and eighth (1994) editions of the widely distributed Handbook on Injectable Drugs. According to the eighth edition of the book, “A 5% solution may be prepared from the 25% product by adding 1 volume of 25% albumin to 4 volumes of sterile water or an infusion solution such as dextrose 5% in water or sodium chloride 0.9%" (6). This section was revised for the 1996 edition (7): 

A 5% solution may be prepared from the 25% product by adding 1 volume of the 25% albumin to 4 volumes of an infusion solution such as dextrose 5% in water or sodium chloride 0.9%. If sterile water for injection is the diluent, the tonicity of the diluted solution must be considered. Substantial reduction in tonicity creates the potential for hemolysis.

We do not know how often pharmacies are using sterile water for injection to dilute 25% albumin solution to 5%. Albumin is still in short supply, so there is a potential for increasingly more pharmacies to incorrectly dilute concentrated albumin solutions. 

To advise the medical community of this potentially serious problem, FDA is taking the following steps: 

  • Recommending to manufacturers of 25% albumin solution that they revise the package inserts to include (1) a warning statement about the risks for potentially fatal hemolysis and acute renal failure when sterile water for injection is used as a diluent and (2) a statement clarifying that 0.9% sodium chloride injection and 5% dextrose injection are the acceptable diluents.         
  • Publishing a “drug warning” in the FDA Medical Bulletin and working with the Centers for Disease Control and Prevention to publish a notice in Morbidity and Mortality Weekly Report about the risks for potentially fatal hemolysis and renal failure from the inappropriate use of sterile water for injection as a diluent for albumin solution used in plasmapheresis.         
  • Working with the American Society of Health-System Pharmacists to revise the albumin monograph in the Handbook on Injectable Drugs (Talley CR, personal communication, 1998 Mar 31). The revision will eliminate ambiguity and state that sterile water for injection should not be used alone to dilute 25% albumin solution, particularly when large volumes of the diluted solution are to be used.         
  • Notifying manufacturers of plasmapheresis devices of this preventable, serious error.
Health care professionals should report any serious adverse events, including cases such as thosewe have described here, to FDA’s MedWatch program (How to report) at 800-FDA-1088 or 800-FDA-0178 (fax) or to the appropriate pharmaceutical manufacturer.
  1. 21 C.F.R. Part 640.82 (1996). Albumin (human): tests on final product.         
  2. Food and Drug Administration. Normal serum albumin (human) and plasma protein fraction (human). Final rule (docket no. 77N-0047). Fed Regist. 1977; 42:27577-8.         
  3. Henry JB. Todd-Sanford-Davidsohn clinical diagnosis and management by laboratory methods. 16th ed. Philadelphia: Saunders; 1979.         
  4. Forte FJ, Caravone D, Coyne MJ. Albumin dilution as a cause of hemolysis during plasmapheresis. Am J Health-Syst Pharm. 1995; 52:207. Letter.         
  5. Steinmuller DR. A dangerous error in the dilution of 25% albumin. N Engl J Med. 1998; 338:1226. Letter.         
  6. Trissel LA. Handbook on injectable drugs. 8th ed. Bethesda, MD: American Society of Hospital Pharmacists; 1994.         
  7. Trissel LA. Handbook on injectable drugs. 9th ed. Bethesda, MD: American Society of Health-System Pharmacists; 1996.
L. Ross Pierce, M.D., Acting Deputy Director Office of Blood Research and Review

Ann Gaines, Ph.D. Frederick Varricchio, M.D., Ph.D., Chief Adverse Events Section Division of Biostatistics and Epidemiology 

Jay Epstein, M.D., Director Office of Blood Research and Review Center for Biologics Evaluation and Research Food and Drug Administration 1401 Rockville Pike Rockville, MD 20852 

A similar letter, written in response to a physician who reported one of the cases, appeared in the April 23, 1998, issue of The New England Journal of Medicine. 

 

Respsonse 

It is common knowledge among pharmacists that large volumes of very hypotonic solutions should not be administered intravenously. Small volumes—not large volumes—of hypotonic solutions can be given intravenously. 

Although FDA recommended in its Federal Register notice that sterile water for injection be used as a diluent for albumin when a lower sodium concentration is indicated (reference 2 in the letter from Dr. Pierce and colleagues), it is undoubtedly suitable only for small volumes of albumin, partial dilutions, combined with other diluents, or nonclinical applications. There are no pharmaceutical products that should be administered in large volumes as very hypotonic solutions. This principle was apparently disregarded in the cases described by Dr. Pierce and colleagues. 

Publication of the letter by Forte et al. (reference 4 in the letter) prompted the inclusion two years ago of a specific notice and reminder about hypotonic solutions in the Handbook on Injectable Drugs. This reminder pointed to the hemolysis that could occur and noted that the tonicity of the final solution must be considered. 

More recently, the January 14, 1998, issue of ISMP Medication Safety Alert (1) had a case report calling attention to the dangers of hypotonic albumin solutions. This alert, by the Institute for Safe Medication Practices, was sent to virtually every hospital pharmacy in the United States and included in the March issue of Hospital Pharmacy (2). I worked with ISMP to produce a follow-up article in the March 25 issue of ISMP Medication Safety Alert (3), also sent to virtually every hospital pharmacy. Collaborations on further notifications continue. 

In the three other cases described by FDA, it appears that the hospital pharmacies relied on old, outdated editions of the Handbook as references. This is a practice that cannot be condoned. The upcoming 10th edition incorporates more than 2000 changes and additions to information in the 9th edition, a process typical for previous editions as well. In the print and electronic versions of the 10th edition, one of these changes will be the greatly strengthened warning about dilution of albumin: 

CAUTION—Substantial reduction in tonicity, creating the potential for fatal hemolysis and acute renal failure, may result from the use of sterile water as a diluent. The hemolysis and acute renal failure that result from the use of a sufficient volume of sterile water as a diluent may be life-threatening.
In addition to this revised warning, a specific notice about albumin dilution appeared in the May 1998 ASHP Newsletter, a boxed warning appeared on the May 1 AJHP table of contents page, and an alert appears on the ASHP World Wide Web site (www.ashp. org).

Pharmacists must use their pharmaceutical education and training, common knowledge, and common sense when interpreting and applying information from the Handbook on Injectable Drugs, or any other reference, to each specific clinical situation. 

  1. Safety briefs. ISMP Med Saf Alert. 1998; 3:Jan 14.         
  2. Cohen MR. Volume limitations for iv drug infusions when sterile water for injection is used as a diluent. Hosp Pharm. 1998; 33:274,277.         
  3. Why you need up-to-date texts! ISMP Med Saf Alert. 1998; 3:Mar 25.
Lawrence A. Trissel, FASHP, Director
Clinical Pharmaceutics Research
Division of Pharmacy, Box 90
The University of Texas
M. D. Anderson Cancer Center
1515 Holcombe Boulevard
Houston, TX 77030

This letter will be published in the May 15, 1998, issue of AJHP with the following citation, "Pierce LR, Gaines A, Epstein J. Hemolysis and renal failure associated with use of sterile water for injection to dilute 25% human albumin solution AJHP 1998; 55:1057-62, 1070. Letters." and was posted 04/21/98.