Catheter Outcomes in the Short-term Inpatient Setting: AControlled Quality Improvement Study Comparing

Citrate and Heparin Lock

Samir Parikh, Ruba Nijmeh, Sheri Van Cleef, Sue Timmerman, Udayan Y. Bhatt, andAnil K. Agarwal

Division of Nephrology, Department of Internal Medicine, The Ohio State University, Columbus, Ohio

ABSTRACT

Heparin and citrate are used as catheter lock solutions toreduce risk of catheter dysfunction and infection in hemodi-alysis. There is a paucity of data comparing these two locksin the short-term, inpatient setting. We compared theefficacy of 2.2% acid citrate dextrose (ACD) versus5000 U ⁄ml heparin as catheter lock in the inpatient setting.The study was conducted at two sites within our system,with ACD locks used at site 1 and heparin locks at site 2.We assessed catheters for catheter dysfunction and infection.Both nontunneled dialysis catheters (NTDC) and tunneleddialysis catheters (TDC) were evaluated. We studied 250

catheters and 139 met inclusion criteria: 90 catheters in theACD group and 49 in the heparin group. ACD had supe-rior outcomes for NTDC; event rate was 0.052 forNTDC ⁄ACD and 0.125 for NTDC ⁄heparin (p = 0.032).There was no difference for TDC. Univariate (odds ratio[OR]: 1.88, confidence interval [CI]: 0.931, 3.82) and multi-variate (OR: 1.35, CI: 0.64, 2.87) analyses demonstrated atrend toward increased odds of event with heparin. Catheterlock with 2.2% ACD has lower risk of catheter dysfunctionas compared with that of 5000 U ⁄ml heparin in the short-term inpatient setting in NTDC and similar risk in TDC.

Background

Maintenance of vascular access patency is critical forpatients on hemodialysis. Despite national guidelinespromoting ‘‘fistula first and catheter last,’’ centralvenous catheters for hemodialysis are used commonlyfor vascular access in prevalent patients and predomi-nantly for those initiating dialysis (1). Late referral ofpatients to nephrologists, lack of placement andmatura-tion of permanent arteriovenous accesses, or emergentinitiation of dialysis in the hospital setting contribute toincreased rates of catheter use for dialysis (2). Unfortu-nately, the use of catheters is complicated by high ratesof dysfunction and infection. Catheter dysfunction ispredominantly a result of intraluminal thrombus forma-tion leading to poor blood flow and inability tomaintaindialysis adequacy with consequent high morbidity,mortality, and costs in terms of increased hospitaliza-tion, thrombolytic use, and catheter exchange. To pre-vent thrombus formation and decrease catheterdysfunction, various catheter locks have been used,

including heparin, citrate, and more recently, recombi-nant tissue plasminogen activator (rt-PA; 3,4).

Heparin has traditionally been the agent of choice tolock hemodialysis catheters. While it is the standard ofpractice, there is little evidence in the literature tosupport its use. Furthermore, the optimal dose has notbeen defined and its use has disadvantages includingsystemic anticoagulation, bleeding, heparin-inducedthrombocytopenia, and biofilm formation (5–8).

Citrate has been used as an anticoagulant for bloodproducts and apheresis for many years and has recentlybeen shown to be both safe and effective as an anticoag-ulant in continuous renal replacement therapies (9,10).Citrate functions as an anticoagulant by chelatingionized calcium in the blood, thus inhibiting the clottingcascade and thus, thrombus formation. Varying concen-trations of citrate have been evaluated with regard tocatheter patency, infection, and other complications par-ticularly in comparison with heparin (11–16). Thesestudies were conducted on stable outpatient hemodialy-sis patients. At present, there are no studies comparingthese catheter locks in the short-term, inpatient setting.

Acute illness requiring hospitalizationmay change thehemodynamic andmetabolicmilieu of patients requiringdialysis. This circumstance differs from the chronic,stable outpatient setting. Acute illness may be associatedwith accentuation of inflammatory state or infection,and this may affect catheter patency and increase riskfor thrombosis and infection. Both nontunneled

Address correspondence to: Anil K. Agarwal, MD, Professorof Internal Medicine and Director, Interventional Nephrol-ogy, The Ohio State University, Columbus, Ohio, e-mail:Anil.agarwal@osumc.edu.

Seminars in Dialysis—Vol 25, No 3 (May–June) 2012 pp. 351–356DOI: 10.1111/j.1525-139X.2011.00966.xª 2011 Wiley Periodicals, Inc.

351

(NTDCs) and tunneled dialysis catheters (TDCs) areused often for acute hemodialysis in the inpatient setting.Maintaining catheter patency in this setting is necessaryto reduce complications, maintain dialysis adequacy,and reduce patientmorbidity.In recent times, there has been a shortage of heparin.

In addition, episodes of heparin contamination in thelast few years have created significant issues and anxietyabout its use in performance of hemodialysis. Conse-quently, our acute inpatient dialysis centers at the OhioState University instituted a policy of change in catheterlock solution in which 2.2% acid citrate dextrose (ACD)replaced heparin as the primary choice for catheter lockfor inpatient hemodialysis. We conducted this compara-tive effectiveness study as a quality assurance project tosystematically study the use of heparin and ACD inshort-term inpatient setting.

Methods

The primary aim of this prospective, cohort study wasto compare the efficacy of 2.2%ACDversus 5000 U ⁄mlheparin as catheter lock solution in both NTDC andTDC in prevention of catheter dysfunction and infectionin the short-term, inpatient setting.

Intervention

The intervention in this study included use of either5000 U ⁄ml heparin or 2.2% ACD as catheter lock.ACD is commercially available as a 2.2% solution madeup of 2.2 g ⁄dl sodium citrate, 0.73 g ⁄dl citric acid, and2.45 g ⁄dl dextrose. This formulation has been used as ananticoagulant in blood collection tubes to preserve redblood cells, as well as for most therapeutic plasmaexchange procedures. We used citrate solution as it wasreadily available at our institution.

Study Groups

This study was conducted at two clinical siteswithin the Ohio State University Health system; inpa-tient acute dialysis main campus (site 1), and inpa-tient acute dialysis East campus (site 2). To maintainconsistent use of the lock solutions, 2.2% ACD lockwas used exclusively at site 1, while 5000 U ⁄ml hepa-rin lock was used exclusively at site 2. The data wereprospectively collected from August 1, 2009 to March1, 2010 (Fig. 1).Both study groups were further classified into NTDC

and TDC groups. The groups were assessed for out-comes based on event rate, which was defined as totalnumber of events (dysfunction and infection) over thetotal number of treatments. Catheter dysfunction wasdefined as poor blood flow (<250 ml ⁄minute), highvenous pressure, having to reverse ports during treat-ment, need for catheter exchange, or catheter thrombo-sis. Catheter thrombosis was defined as the inability todraw or flush one or two ports, aspiration of clots, orneed to use rt-PA. Catheter infection was defined as doc-umented bacteremia attributed to the catheter.

Inclusion and Exclusion Criteria

We included all NTDCs and TDCs placed for perfor-mance of intermittent hemodialysis during the studyperiod. All catheters were inserted by experiencednephrology fellows, nephrology attending physicians, orby interventional radiologists using standard Seldingertechnique under sterile conditions. Ultrasound and ⁄orfluoroscopy guidance were used for all insertions. Stan-dard antiseptic technique was used by all staff whenaccessing the catheters and instilling catheter locks.We excluded all catheters inserted outside our data

collection period, with uncertain date of insertion, usedat any point for purpose other than hemodialysis, cathe-ters in which both heparin and ACDwere used as locks,and those used for one treatment only or for continuoustherapies.

Statistical Methods

Baseline characteristics were compared (Table 1).Categorical variables in the two groups were analyzedusing chi-square analysis. Continuous variables werecompared using Student’s t-test. As multiple treatmentswere provided per subject, generalized estimating equa-tion (GEE) analysis was used to account for patient-specific and exposure-specific effects. Modeling wasused to estimate the odds of catheter dysfunction in theheparin group relative to the citrate group. Univariate

Site # 1Citrate OnlyN = 90Tx # = 418

Site # 2Heparin OnlyN = 49Tx # = 186

Excluded CathetersN = 111Criteria above

Total Catheters involved in study N = 139Tx # = 604

TDCN = 27Tx # = 148

NTDCN = 63Tx # = 270

TDCN = 24Tx # = 106

NTDCN = 25Tx # = 80

Total Number of Catheters (TDC and NTDC)N = 250

Fig. 1. Study design chart.

TABLE 1. Baseline characteristics

Heparin ACD p-value

Number of catheters 49 90 –Age 61.14 � 2.12 61.08 � 1.67 0.983Gender (male %) 42.9 55.6 0.152African American (%) 85.7 28.9 <0.001Catheter location 0.228Right IJ (%) 38.8 53.9 –Left IJ (%) 44.9 34.8 –Femoral (%) 16.3 11.2 –

ACD, acid citrate dextrose; IJ, internal jugular.

352 Parikh et al.

and multivariate models were created. Variablesincluded in the multivariate model included catheterlock type (heparin or citrate), age, gender, race, proce-dure center, catheter type (tunneled or nontunneled),and location of catheter placement. All analyses wereperformed using STATA11 (Austin, TX,USA).

Results

We evaluated 250 dialysis catheters placed during thestudy period, of which 139 catheters met inclusion crite-ria. This amounted to 1172 catheter days and 604 totaltreatments. The ACD group had 90 dialysis catheters:63 NTDCs and 27 TDCs. The heparin group had 49dialysis catheters: 24 NTDCs and 25 TDCs. The totalnumber of catheter days and treatments for the citrategroup was 746 and 418, respectively. The heparin grouphad 426 catheter days and 186 treatments (Table 2).

For both TDCs and NTDCs, the overall event ratesand event rates for catheter dysfunction and clottingonly were compared and a statistically significant reduc-tion in the number of events in the ACD group versusthe heparin group for NTDCs was observed. The eventrates were similar for TDCs (Figs 2 and 3). A subgroupanalysis was then preformed for African Americanpatients with NTDC only and the results similarlyfavored theACDgroup (Fig. 4).

Considering all catheter types in the univariate analy-sis, the odds ratio (OR) of a catheter-related event in theheparin group was 1.88 (0.931, 3.82) relative to the ACDgroup.Multivariate analysis was used to adjust for cath-eter lock type, age, gender, race, procedure center, cathe-ter type, and location of catheter placement. Afteradjusting for these variables, the OR was 1.35 (0.64,2.87) for catheter-related events in the heparin grouprelative to the ACD group (Fig. 5). As demonstrated,there was a nonsignificant trend toward an increase incatheter-related events in the heparin group relative totheACDgroup.

The results were then stratified by catheter type. Theevent rate for the NTDC-ACD group was 0.052 with270 treatments and 447 catheter days. This group had 2episodes of infection and only 12 episodes of catheterdysfunction ⁄ thrombosis in a total of 63 catheters. Theevent rate for theNTDC-heparin groupwas 0.125 for 80treatments and 155 catheter days. This group had noincidence of infection, but there were 10 episodes ofcatheter dysfunction ⁄ thrombosis in 24 catheters. Inunadjusted analysis, the OR was 2.55 (1.08, 6.04;

TABLE 2. Event rates per treatment: acid citrate dextrose (ACD)

versus heparin in nontunneled dialysis catheters (NTDC) and

tunneled dialysis catheters (TDC)

Catheter typeand lock (n)

Totalcatheterdays

Totalnumber oftreatments

Catheterinfection

Catheterdysfunction

Eventrate

ACD-NTDC(63)

447 270 2 12 0.052

Heparin-NTDC(25)

155 80 0 10 0.125

ACD-TDC (27) 299 148 0 8 0.054Heparin-TDC(25)

271 106 1 5 0.06

Fig. 2. Overall event rates for acid citrate dextrose and heparin

in tunneled and nontunneled dialysis catheters.

Fig. 4. Subgroup analysis comparing heparin with acid citrate

dextrose in nontunneled dialysis catheters in African Americans.

Fig. 3. Clotting and catheter dysfunction in tunneled and

nontunneled dialysis catheters between heparin and acid citrate

dextrose.

SHORT-TERM CATHETER OUTCOMES WITH CITRATE VS. HEPARIN LOCK 353

p = 0.032) for catheter-related events in the heparingroup relative to the ACD group for NTDC (Fig. 6).There were more infections in the ACD group, but theincidence of infection in all groupswas very low and thuswe were unable to assess the difference between the twogroups (Table 3).The event rate for the TDC-citrate group was 0.054

with 148 treatments and 299 catheter days. This grouphad 8 episodes of catheter dysfunction ⁄ thrombosis withno infection in 27 catheters. In comparison, the TDCheparin group had an event rate of 0.06 with 106 treat-ments and 271 catheter days. This group had 5 episodesof catheter dysfunction ⁄ thrombosis and 1 episode of

infection in 25 catheters. In univariate analysis, the ORwas 1.42 (0.44, 4.58) for catheter-related events in theheparin group relative to the ACD group for TDC. Amultivariate adjustment was not attempted because ofsmall sample size.

Discussion

Heparin has been used as a catheter lock in dialysiscatheters for many decades. While there are frequentproblems associated with heparin, its use has not beenstudied systematically. One study demonstrated systemicanticoagulation with heparin locks at 5000 Uml and10,000 U ⁄ml with partial thromboplastin times twicethe normal limit at 10 minutes postinstillation and stillelevated at 1 hour (6). In vitro studies have demon-strated that even when catheters are locked exactly tothe recommended volume, 15–20% of the volume leakssystemically (5). In another study, less systemic anticoag-ulationwas foundwith citrate lock comparedwith hepa-rin (12).The recent shortages in heparin supply along with

issues with contamination led to a policy change at ourinstitution to ACD-only catheter locks for hemodialysiscatheters during the acute short-term inpatient use. Forreasons of paucity of data regarding outcomes of suchan intervention, we secured this opportunity to evaluatewhether a change to ACD-only lock would be appropri-ate. The results of our study demonstrated 2.2% ACDto be equal to and possibly more efficacious than5000 U ⁄ml heparin as catheter lock for short-term inpa-tient dialysis.Historically, citrate use for catheter lock has been

reserved for patients with heparin allergy or history ofheparin-induced thrombocytopenia. Citrate, at bothhigh and low concentrations, has been compared withheparin as catheter lock in the chronic, outpatientsetting. A majority of these studies demonstrated equiv-alence between the two agents with regard to catheterdysfunction, need for rt-PA, catheter exchange, andaccess-related hospitalizations. A recent randomizedstudy evaluated 61 chronic hemodialysis patients withcuffed, tunneled dialysis catheters to tri-sodium citrate(TSC; 4%) or 5000 U ⁄ml heparin. They found compa-rable events of catheter dysfunction in 13 ⁄32 (41%)versus 12 ⁄29 (41%) in the respective groups (15). Cathe-ter-related bacteremia was also similar between the twogroups (15). Another study found 30%TSC to be equiv-alent to 5000 U ⁄ml heparin when used as catheter lockin preventing thrombosis, use of rt-PA, and access-related hospitalizations (13). Only one study comparinghigh-concentration citrate (46.7% TSC) with heparin5% for catheter lock found, as a secondary endpoint,increased incidence of catheter thrombosis and moreadverse events with the high-dose citrate group (16).An ideal catheter lock will prevent infectious compli-

cations.Heparin lock, however, does not prevent biofilmformation in the presence of Staphylococcus aureus andin one study, higher doses of heparin promoted biofilmformation of S. aureus (7,8). Citrate has been shownto prevent biofilm formation and has demonstrated

Fig. 6. Odds ratio of catheter dysfunction in nontunneled dialy-

sis catheters: unadjusted analysis.

TABLE 3. Infection rate: acid citrate dextrose (ACD) versus

heparin in both nontunneled dialysis catheters (NTDC) and tunneled

dialysis catheters (TDC)

ACD Heparin p-value

NTDC 0.007 0.00 Not calculatedTDC 0.0 0.009 Not calculated

Fig. 5. Multivariate analysis: heparin versus citrate for all treat-

ments in both tunneled and nontunneled dialysis catheters.

354 Parikh et al.

antimicrobial activity in vitro with several studies report-ing reduced catheter-related bacteremia with citratecompared with heparin (3). In a randomized controlledtrial, 30% TSC demonstrated decreased biofilm forma-tion and less catheter colonization when compared with5% heparin as catheter lock (17). Furthermore, 30%TSC was shown to have better antimicrobial activityand less catheter-related bacteremia compared with5000 U ⁄ml heparin (11,18).

High-concentration citrate solutions can leak intothe systemic circulation, with risk of chelation of cal-cium. The consequence of this includes metallic taste,paresthesias, and arrhythmias. The potential risk offatal cardiac arrest from high-concentration citrate ledto its removal from the market in the United States(19,20). Low-concentration citrate, however, hasshown a better safety profile and has not been foundto cause systemic anticoagulation or hypocalcemia.There have been several studies demonstrating the effi-cacy and safety of low-concentration citrate (4%) ascatheter lock. These studies have demonstrated equiv-alent or fewer complications with low-concentrationcitrate compared with varying doses of heparin.

rt-PA is significantly more expensive than heparinand has traditionally been used as a rescue lock torestore patency of dysfunctional catheters despite use ofheparin lock. A recent multicenter, double-blindrandomized control trial found replacing heparin lockwith once-weekly rt-PA reduced incidence of catheterdysfunction and catheter-related bacteremia (4). Theoutcomes were encouraging, but of concern was the inci-dence of adverse events. Seventy percent of patients inthe rt-PA group and 68.7% in the heparin group had anadverse event. As mentioned before, leak of catheterlock and systemic anticoagulation have been demon-strated with heparin locks. The same concern remainswith rt-PA use. Cost also remains an issue as rt-PA is sig-nificantlymore expensive than the heparin.

With any lock solution over long-term, cost mustbe considered. While heparin costs much less than rt-PA, it is much more expensive than citrate. A pro-spective trial assessing the efficacy, cost, and safety of4% TSC versus heparin found equivalent or betteroutcomes with TSC with regard to need for catheterexchange, rt-PA use, and access-related hospitaliza-tions compared with heparin (14). At our institution,a 500-ml bag of citrate costs $5.18 with only 5 mlused per treatment and its estimated cost is $0.05 percatheter lock. In contrast, 5000 U ⁄ml heparin costs$1.82 per treatment which, while still inexpensive, is36 times more expensive than citrate. Considering ourdata and the recent literature, citrate in low concentra-tions may be a good alternative for heparin at lesscost and with a better safety profile.

Our study has important limitations. First, it is aparallel cohort study involving two locations, with a dis-crepancy in the African American population betweenthe two sites. However, the rest of the baseline character-istics were similar. The difference in settings is importantin establishing generalizability of our findings and mayin fact be advantageous. Second, heparin locks wereused only at site 2 while ACD locks were used at site 1.

One could argue that the differences in demographicsand catheter use practices at the two centers wereresponsible for the difference, not the lock solution.However, the two sites follow identical healthcare prac-tices and the personnel and policies related to cathetercare belong to the same healthcare system. Nevertheless,this possibility cannot be ruled out entirely. Third, oursample size is small because of limited enrollment period,as both institutions were to convert completely to citrateonly by March 1, 2010. However, we had adequatestatistical power to detect a difference between the out-comes of NTDC. Despite these shortcomings, our find-ings are consistent with previous studies comparingcitrate with heparin in the chronic outpatient dialysissetting. A larger randomized, controlled study, prefera-bly atmultiple centers, would be necessary to verify thesefindings.

Conclusion

Our study showed that 2.2% ACD has advantagesover 5000 U ⁄ml heparin as catheter lock for short-terminpatient dialysis. The results from this prospective,cohort study suggest better outcomes with ACD inNTDC in the inpatient setting for short-term hemodialy-sis. There was no significant difference in catheteroutcomes between the two groups for TDC in this set-ting. Our data suggest that 2.2%ACD as a catheter lockis superior to 5000 U ⁄ml heparin in preventing dysfunc-tion forNTDCs and is similar to it for TDCs.

References

1. NKF-K ⁄ DOQI Clinical Practice Guidelines for Vascular Access.Update 2006. Available at http://www.kidney.org/professionals/KDOQI/guideline_upHD_PD_VA/va_guide2.htm, accessed May 12,2011

2. Rayner HC, Besarab A, Brown WW, Disney A, Saito A, Pisoni RL:Vascular access results from the Dialysis Outcomes and Practice Pat-terns Study (DOPPS): performance against Kidney Disease OutcomesQuality Initiative (K ⁄ DOQI) Clinical Practice Guidelines. Am J KidneyDis 44:S22–S26, 2004

3. Moran JE, Ash S, Clinical Practice Committee: Locking solutions forhemodialysis catheters; heparin and citrate – a Position Paper byASDIN. Semin Dial 21:490–492, 2008

4. Hemmelgarn BR, Moist LM, Lok CE, Tonelli M, Manns BJ, HoldenRM, LeBlanc M, Faris P, Barre P, Zhang J, Scott-Douglas N: Preven-tion of dialysis catheter malfunction with recombinant tissue plasmino-gen activator. N Engl J Med 364:303–312, 2011

5. Sungur M, Eryuksell E, Yavas S, Bihorac A, Layon J, Caruso L: Exit ofcatheter lock solutions from double lumen acute hemodialysis catheters:an in vitro study. Nephrol Dial Transplant 22:3533–3537, 2007

6. Karaaslan H, Peyronnet P, Benevent D, Lagarde C, Rince M, Leroux-Robert C: Risk of heparin lock-related bleeding when using indwellingvenous catheters for hemodialysis. Nephrol Dial Transplant 16:2072–2074, 2001

7. Shanks R, Sargent J, Martinez R, Graber M, O’Toole G: Catheter locksolutions influence staphylococcal biofilm formation on abiotic surfaces.Nephrol Dial Transplant 21:2247–2255, 2006

8. Shanks R, Donegan N, Graber M, Buckingham S, Zegans M, CheungA, O’Toole G: Heparin stimulates staphylococcal biofilm formation.Infect Immun 73:4596–4606, 2005

9. Burry L, Tung D, Hallet D, Bailie T, Carvalbana V, Lee D, RamganeshS, Richardson R, Mehta S, Lapinsky S: Regional citrate anticoagulationfor Prismaflex continuous renal replacement therapy. Ann Pharmacol43:1419–1425, 2009

10. Oudemans-van Straaten H: Citrate anticoagulation for continuous renalreplacement therapy in the critically ill. Blood Purif 29:191–196,2010

SHORT-TERM CATHETER OUTCOMES WITH CITRATE VS. HEPARIN LOCK 355

11. Weijmer MC, Van den Dorpel MA, Van de Ven PJG, ter Wee PM, vanGeelen JACA, Groeneveld JO, van Jaarsveld BC, Koopmans MG,lePoole CY, Schrander-Van der Meer AM, Siegert CEH, Stas KJF forthe CITRATE study group. Randomized, clinical trial comparison oftrisodium citrate 30% and heparin as catheter-locking solution in he-modialysis patients. J Am Soc Nephrol 16:2769–2777, 2005

12. Bayes B, Bonal J, Romero R: Sodium citrate for filling hemodialysiscatheters. Nephrol Dial Transplant 14:2532–2533, 1999

13. Stas K, Vanwalleghem J, De Moor B, Keuleers H: Trisodium citrate30% vs. heparin 5% as catheter lock in the interdialytic period in twinor double-lumen hemodialysis catheter for intermittent hemodialysis.Nephrol Dial Transplant 16:1521–1522, 2001

14. Lok CE, Appleton D, Bhola C, Khoo B, Richardson RM: Trisodiumcitrate 4% – an alternative to heparin capping of hemodialysis catheters.Nephrol Dial Transplant 22:447–483, 2007

15. MacRae J, Dojcinovic I, Djurdjev O, Jung B, Shalansky S, Levin A,Kiali M: Citrate 4% versus Heparin and the Reduction of ThrombosisStudy (CHARTS). Clin J Am Soc Nephrol 3:369–374, 2008

16. Power A, Duncan N, Singh S, Brown W, Dalby E, Edwards C, LynchK, Prout V, Caims T, Griffith M, McLean A, Palmer A, Taube D: Cen-tral venous catheter locks: heparin versus citrate. Am J Kidney Dis53:1034–1041, 2009

17. Bosma JW, Siegert CE, Peerbooms PG, Weijmer MC: Reduction ofbiofilm formation with TSC in hemodialysis catheters: a randomizedcontrolled trial. Nephrol Dial Transplant 25:1213–1217, 2010

18. Weijmer MC, Debets-Ossenkopo YJ, Van De Vandervoort FJ, ter WeePM: Superior antimicrobial activity of trisodium citrate over heparinfor catheter locking. Nephrol Dial Transplant 17:2189–2195, 2002

19. Polaschegg HD, Sodemann K: Risks related to catheter lockingsolutions containing concentrated citrate. Nephrol Dial Transplant18:2688–2690, 2003

20. FDA: Food and Drug Administration. FDA Issues Warning on Tricitra-sol Dialysis Catheter Anticoagulant. FDA Talk Paper, 2000:T00-16 U.S.Department of Health and Human Services. Public Health Service 5600Fishers Lane Rockville, MD 20857

356 Parikh et al.