Printer Friendly

Clinical Laboratory Quality Practices When Hemolysis Occurs.

Enormous advances in science and technology have characterized the 20th and 21st centuries, which have transformed the practice of clinical medicine so that history-taking and physical examination have given way increasingly to a practice of medicine dominated by the use of medical technology, and in particular, laboratory testing. (1) Laboratory testing requires obtaining a specimen from patients, making these specimens ready for measurement, and quantifying and reporting results for the analytes of interest by complex instrumentation. Phlebotomy is the method of choice for obtaining the blood specimen, and, for most chemical measurements, the analyte of interest is separated from cellular elements and is then quantified in serum or plasma. Whole-blood measurements have come into vogue when speed of analysis is essential, and these specimens are analyzed without taking time to separate serum or plasma from the cellular elements. Although instrumentation improves clinical laboratory measurement accuracy, interferences frequently occur in specimens, of which, the most common is hemolysis. (2) Regulatory requirements as well as good laboratory practices demand that personnel identify common and critical quality issues in their daily practice and that improvement ensues.

Hemolysis interferes with chemistry tests in a number of ways, with 2 of them being the most important. Most laboratory tests in chemistry are still based on the measurement of light that passes through a specimen. Because of the red color of hemoglobin, hemolysis can interfere with the absorption of light as it passes through a specimen, with the characteristics of interference being dependent on both the method and the instrument. In spectrophotometric assays, oxyhemoglobin has an absorption peak between 531 and 543 nm, but the absorbance curve is relatively broad so hemolysis interferes across many wavelengths used for analyte quantification. (3) The second major way that hemolysis interferes with laboratory tests is by liberating into the serum or plasma analytes that are in high concentration within the red blood cell and are far less concentrated in serum or plasma, thereby giving spuriously increased results. Examples of such spuriously elevated results for analytes are potassium and lactate dehydrogenase (LD). (4) Some measurements, such as those made for electrolytes using whole blood, even ignore the occurrence of hemolysis because there is no practical way to measure its occurrence or effects. (5) Of less importance are analytes that are in low concentration in red blood cells and in high concentration in serum, such as glucose. (4)

Although hemolysis is a general problem in all laboratories, recent studies have indicated that the frequency of hemolysis is highly variable among institutions because it is influenced by the training and competency of the personnel obtaining the specimens, the equipment used to procure the specimens, the location where the patient is seen, the manner in which specimens are procured, and what has been done to reduce the rate of hemolyzed specimens. Although laboratory personnel have tried to reduce the frequency of hemolyzed specimens, there are no recommendations from interlaboratory studies on the potential success of these actions. In addition, how clinical laboratory personnel handle hemolyzed specimens is highly variable and is dependent on individual laboratory policies and procedures, instrument manufacturers' recommendations, the analyte being measured, and the specimen type. We studied clinical laboratory and institutional policies for detecting and handling hemolyzed specimens and the actions taken to reduce hemolysis. We conclude that laboratorians must improve hemolysis practices and offer suggestions for improvement.

MATERIALS AND METHODS

Members of the College of American Pathologists (CAP) Quality Practices Committee developed an online questionnaire in the usual CAP Q-Probes format, which was sent to participants in the 2011 CAP C1, C3, C3X, CZ, CZX, CZ2, and CZ2X chemistry surveys. One reminder notice was sent to nonrespondents, encouraging participants to complete their questionnaires and return them to CAP's Q-Probes committee staff, who then tabulated the results according to previously established Q-Probes procedures. (6) The questionnaire contained 28 multiple-choice questions, with "other, please list" a choice in 7 questions. If participants did not answer a question, they were still included in the responses of the questions they answered. Those few participants who responded that they did not know their laboratories' practices for a specific question were excluded from the database for that specific question.

RESULTS

A total of 846 customers of the CAP Chemistry Surveys responded to the questionnaire for a 24% response rate. Figure 1 shows laboratory practices for 3 different analytes: potassium, glucose, and LD. The percentage of clinical laboratories with written policies for rejection varied from 85% for potassium to 69% for LD to 55% for glucose. A few participants standardized and correlated their hemolysis findings between their primary and second chemistry analyzers (46% for potassium and 40% for glucose and LD), with potassium analytes receiving the most attention.

Figure 2 shows the percentages of 694 participants who tried to validate their manufacturers' findings on the influence of hemolysis on measurements for potassium, glucose, LD, or any analyte. Of the participants, 30% attempted to validate the interference for at least one analyte, with most attempting to validate the effect of hemolysis on potassium measurements (29%); fewer attempted to validate hemolysis on glucose measurements (24%), and attempting to validate the effect of hemolysis on LD measurements (23%) was attempted by the fewest participants. Essentially all participants who attempted validation indicated they were able to confirm their manufacturers' findings.

Figure 3 describes actions taken by 692 participants during the past year to decrease hemolysis. Forty-nine percent of participants monitored hemolysis and took corrective action to decrease the percentage of hemolyzed specimens received in their laboratories. Of these, 47% of participants systematically and regularly monitored the percentage of hemolyzed specimens as part of their performance improvement plan. Most of these participants (34%) monitored the percentage of hemolyzed specimens systematically and regularly throughout the entire institution, whereas 13% had similar improvement practices, but only for specimens originating from specific locations, such as the Emergency Department (ED).

Figure 4 provides the 806 specific actions that participants took to decrease the rate of hemolysis in 337 clinical laboratories monitoring hemolysis. On average, 2.4 different actions were taken throughout the year in each laboratory as part of its performance improvement plan to reduce the percentage of tests with hemolysis. Collecting data and distributing it to leadership in phlebotomy locations (57%), troubleshooting individual specimens with phlebotomists (55%), providing ongoing training sessions for phlebotomists (53%), establishing a quality improvement team between the clinical laboratories and offending locations (37%), and providing information for physicians (19%) were the 5 most commonly used interdepartmental improvement practices. There were 37 other practices, such as replacing nurse phlebotomists with laboratory personnel, restricting blood collection through indwelling intravenous catheters, or changing blood collection tubes, each implemented in a small number of institutions (5 or 6).

Most participants (71% of 717 participants) agreed about their lack of success in decreasing the percentage of hemolyzed specimens. Of these, 50% said they continued to try, but had made slow progress; 10% had tried, but now, did not know what to do; 8% continued to try to reduce hemolysis, but nothing seemed to help; 2% had become frustrated and had given up trying to reduce hemolysis; and less than 1.0% said they now ignore hemolysis (Figure 5).

Approximately 60% of clinical laboratories use the same scale for hemolysis, lipemia, and icterus for measurements of potassium (n = 627), glucose (n = 662), and LD (n = 684) as shown in Figure 6.

COMMENT

Clinical laboratory accrediting agencies are required to monitor and enforce important clinical laboratory quality practices developed and described under the Clinical Laboratory Improvement Amendments of 1988 (CLIA '88) by the Centers for Medicare & Medicaid Services. Requirements listed in the largest accrediting program, the CAP Laboratory Accreditation Program, are that each clinical laboratory must reconfirm manufacturers' claims of important test characteristics, such as analytic measurement range, linearity, and clinical measurement range for all quantitative measurements, and that analytic quality of all analytes be accessed by proficiency testing. (7) Failure to reconfirm the analytic measurement range, linearity, and the clinical measurement range appropriately will result in citations upon inspection, and if the citations are extensive and serious, the clinical laboratory may be closed. Proficiency testing of complex analytes, such as potassium, LD, and glucose, are conducted using unknown samples 15 times yearly. Clinical laboratories whose proficiency results consistently exceed specific analytic criteria limits for an analyte may lose accreditation for the inaccurate measurements and would need to cease and desist from testing this analyte.

The CAP proficiency testing limits of acceptability are set at [+ or -] 0.5 mEq/L of the target value for potassium, [+ or -] 6 mg/dL or [+ or -] 10% of the target value for glucose, and [+ or -] 20% of the target value for LD. (8) For the CAP Calibration Verification/ Linearity Surveys, these limits are set at a total error of [+ or -] 10% for potassium, [+ or -] 2 mg/dL or [+ or -] 6.0% for glucose, and [+ or -] 20% for LD. (9) For analytes such as potassium, LD, or glucose, a hospital whose clinical laboratory is unable to perform these measurements because of the associated stiff penalties for poor analytic performance may find it difficult to remain open.

Because the concentration of potassium in red blood cells is 23 times higher than it is in serum and LD in red blood cells is 160 times higher than it is in serum, Laessig pointed out almost 50 years ago that a 1% lysis of red blood cells caused an increase of 55% for potassium and 272% for LD. (4) Higher rates of hemolysis will lead to even higher elevations of potassium and LD. Some analytes, such as glucose, are in lower concentration in red blood cells than they are in serum, and when a similar rate of hemolysis occurs, a 5% decrease in glucose will be found. The effect of 1% hemolysis described by Laessig markedly exceeds the allowable error of potassium and LD and approaches the allowable error for glucose in the CAP Proficiency and Calibration Verification/Linearity Programs. Therefore, it is not surprising that frequent incidents occur throughout almost all clinical laboratories when analytic errors in testing caused by hemolysis far exceed these guidelines developed under CLIA '88 for acceptable proficiency and linearity testing. Despite these widely discrepant results caused by hemolysis, improvement is not mandated by governmental and accrediting agencies.

Throughout the study, participants had more concerns for potassium than for glucose or LD measurements, as manifested by a higher percentage of activities performed for this analyte. Such activities included a greater percentage of participants with hemolysis specimen rejection procedures for potassium (85%) than for LD (69%) or glucose (55%), and standardized hemolysis reporting between the primary and the second analyzer (46%) for potassium compared with LD or glucose (40%). Although a few participants had specific procedures for correlating and expressing hemolysis results among their major chemistry analyzers, it is common practice for laboratorians to choose the same brand and model of analyzer as both their primary and secondary instruments. Only recently have instrument manufacturers begun offering instruments that customize the reporting of hemolysis levels on an analyte-by-analyte basis, and some have even received US Food and Drug Administration approval for their methods. Some participants may not have correlated their response between their 2 analyzers because they may have relied inappropriately on a visual, rather than an automated, evaluation of hemolysis. (5)

Despite the criticality of the effect of hemolysis on clinical laboratory tests, we found it is an uncommon practice to confirm the manufacturers' stated claims of the interferences of hemolysis on potassium, glucose, and LD, because only up to 30% of the participants tried to do so. Almost all who tried to confirm hemolysis chose to study potassium, with fewer studying glucose, and the fewest choosing LD. Essentially all participants in our study who tried to confirm their manufacturers' claims were able to do so. Interference of hemolysis on various analytes has been found to be method specific, and with the abilities of some of the newer instruments to routinely incorporate measurements of hemolysis as well as to express the effect about the degree of hemolysis on an analyte-by-analyte basis, significant opportunities now exist for improvement. Despite the widely discrepant results caused by hemolysis and the requirement of accreditation agencies, such as the CAP, that clinical laboratory staff must study and provide interference studies to inspectors, only 30% of our participants had validated manufacturers' hemolysis studies and would be able to provide data if challenged by an inspector.

Although most participants have tried to reduce hemolysis, the responses of participants attest to the difficulties in reducing hemolysis. Half of the participants continued to try, even though they have made slow progress; 10% were concerned but did not know how to reduce it further; 8% continued to try, but nothing seemed to work; and 2% even have given up because nothing worked. In our experience, part of the reason for the frustration of clinical laboratory personnel is that in an increasing number of clinical laboratories, the clinical laboratory staff no longer provides phlebotomy for most of the patients because nursing personnel are providing an increasing percentage of the phlebotomies. It also is common knowledge among laboratorians that collection techniques are the origin of the hemolysis in locations such as the ED, despite nursing personnel vigorous denials. (10) Because nursing staff do not directly report to clinical laboratory staff, in many circumstances, it is difficult to require that these staff members improve their specimen-collection skills. Many of the most important reasons for continued hemolysis from phlebotomy include pressure differences and needle size, (11,12) phlebotomy sites below the antecubital fossa, (13,14) size of collection tubes, (12,15) difficulty of blood drawing, (12,16) collections through indwelling catheters (12,17,18) and the use of a vacutainer system. (18,19)

In most clinical laboratories, clinical laboratory personnel had multiple actions ongoing yearly to decrease hemolysis by collecting and sharing hemolysis data, educating hospital personnel about hemolysis, repeatedly training personnel who provide specimens with hemolysis, and developing performance improvement teams between the clinical laboratory and staff in problem locations. The many activities represent a huge outlay of resources in an attempt to decrease the rate of hemolysis, with many imaginative solutions developed within hospitals. Lippi et al (2) have categorized 21 different causes of in vitro hemolysis in a review. Education was shown to reduce the rate of hemolysis, from 19.8% to 4.9%, in one ED, where 25 different phlebotomy characteristics were employed, including publicizing hemolysis rates according to the name of the phlebotomist. (20)

Clinical practice guidelines for phlebotomy in the ED have recently been published by expert panels that evaluated published and unpublished data. They classified data by practices that had a high degree of clinical certainty as being beneficial, those that had moderate clinical certainty and that were likely to be beneficial, as those that had limited or unknown effectiveness, and those that were not recommended for practice. They only recommended those practices that had a high or moderate degree of certainty in being able to reduce hemolysis. (14,21) Variables considered included evidence of the use of (1) phlebotomists versus nursing or ED staff, (2) straight-needle venipunctures versus existing intravenous catheters, (3) syringes versus vacuum tubes when intravenous catheters are in place, (4) collections at an antecubital site versus distal sites, (5) larger (<21-gauge) needles or catheters versus smaller (>21-gauge) needles or catheters, (6) vacuum blood collection tubes versus full vacuum tubes, (7) limiting tourniquet times to [less than or equal to] 1 minute versus times longer than 1 minute, (8) education, (9) monitoring and feedback, and (10) specimen transport. The recommendations made by one or both of these groups (2,20) were that hemolysis was less likely when blood was drawn from the antecubital fossa or when direct venipunctures with straight needles were used, rather than when blood collections occurred through intravenous catheters. Other recommendations included drawing blood through needleless connectors, which did not increase hemolysis, and using low-vacuum blood collection tubes, which resulted in less hemolysis than full vacuum tubes.

Performance improvement techniques involve FOCUS PDCA (find, organize, clarify, understand, select, plan, do, check, act), and we would suggest an interdisciplinary team, similar to what was organized in 37% of the clinical laboratories, as the most important action to take for improvement. Such improvement teams should focus on many of the 21 variables described by Lippi et al (2) or the 25 variables described by Ong et al (20) when beginning a systematic study to reduce hemolysis.

In our experience, the most commonly used corrective action of passing hemolysis data to either the hospital leadership where the hemolysis occurred or to physicians is the action that, if carried out alone, is the least likely to cause improvement because the A or "act" of PDCA is thereby least likely to occur. Troubleshooting outliers and educating phlebotomists on appropriate blood collection techniques or retraining poorly performing phlebotomists are appropriate actions to be used alone or by an interdisciplinary team. In a recent study, Hawkins (22) found that physicians, as well as clinical laboratory staff, have been puzzled by the meaning and significance of hemolysis. This cognitive error found by Hawkins may explain why hemolysis has been overlooked for more than 50 years and why performance improvement practices that decrease hemolysis have been so difficult to implement. Because approximately 60% of participants monitor and report rates of icterus and lipemia in a manner similar to rates of hemolysis, similar opportunities for improvement exist with these interferences as well.

The clinical laboratory has a critical role in health care. One widely quoted estimate of that criticality is that 70% of all medical decisions are based on laboratory results. The fiscal consequences of rejected or poor-quality blood specimens were evaluated (23) in 7 hospitals throughout the world using an elaborate model based on institutional and departmental fiscal information and clinical practice data. Average cost savings of more $1.1 million for a 650-bed hospital was predicted if preanalytic specimen errors, such as hemolysis, could be eliminated. (23)

We conclude that hemolysis remains one of the last frontiers in laboratory specimen quality and suggest that clinical laboratory staff choose one of the newer instruments that measures hemolysis, quantify and report whether hemolysis is present, and then report the influence of hemolysis to clinicians on an analyte-by-analyte basis. Laboratorians should also validate the effect of hemolysis on all analytes measured, use the same hemolysis scales for all analytes and instruments, and develop policies and procedures for hemolysis for all methods performed. It is important that laboratorians work with others throughout their organizations in interdepartmental teams to decrease the percentage of hemolyzed specimens by focusing on the process of phlebotomy. We also suggest that accrediting agencies develop and enforce requirements of good laboratory practices that require laboratorians to evaluate the effect of interference, such as hemolysis, lipemia, and icterus, on a method-by-method basis. Lastly, it is essential that instrument manufacturers begin to focus on how hemolysis or hemoglobin can be measured when it interferes with an analyte that is measured using their whole-blood analyzer.

References

(1.) Burke, MD. Laboratory medicine in the 21st Century. Am J Clin Pathol. 2000;114(6):841-846.

(2.) Lippi G, Blanckaert N, Bonini P, et al. Haemolysis: an overview of the leading cause of unsuitable specimens in clinical laboratories. Clin Chem Lab Med. 2008;46(6):764-772.

(3.) Makarem, A. Hemoglobins, myoglobins, and haptoglobins. In: Henry RJ, Cannon DC, Winkelman JW, eds. Clinical Chemistry: Principles and Techniques. Hagerstown, MD: Harper & Row; 1966:1125-1127.

(4.) Statland BE, Winkel P. Sources of variation in laboratory measurements. In: Henry JB, ed. Clinical Diagnosis and Management by Laboratory Methods. 16th ed. Philadelphia, PA: WB Saunders; 1979:3-28.

(5.) Howanitz PJ, Lehman CM, Jones BA, Meier FA, Horowitz GL. Practices identifying and rejecting hemolyzed specimens are highly variable in clinical laboratories. Arch Pathol Lab Med. In press.

(6.) Howanitz PJ. Quality assurance measurements in department of pathology and laboratory medicine. Arch Pathol Lab Med. 1990;114(11):1131-1135.

(7.) College of American Pathologists. All Common Checklist. Northfield, IL: College of American Pathologists; 2012.

(8.) College of American Pathologists. CC Chemistry/Therapeutic Drug Monitoring. College of American Pathologists Northfield, iL. 2012.

(9.) College of American Pathologists. LN2-B Calibration Verification/Linearity Chemistry/Lipid/Enzyme Survey. Northfield, IL: College of American Pathologists; 2012.

(10.) Pretlow L, Gandy T, Leibach EK, Russell B, Kray B. A quality improvement cycle: hemolyzed specimens in the emergency department. Clin Lab Sci. 2008; 21(4):219-224.

(11.) Sharp MK, Mohammad SF. Scaling of hemolysis in needles and catheters. Ann Biomed Eng. 1998;26(5):788-797.

(12.) Dugan L, Leech L, Speroni KG, Corriher J. Factors affecting hemolysis rates in blood samples drawn from newly placed IV sites in the emergency department. J Emerg Nurs. 2005(4);31:338-345.

(13.) Fang L, Fang SH, Chung YH, Chien ST. Collecting factors related to the haemolysis of blood specimens. J Clin Nurs. 2009:17(17):2343-2351.

(14.) Heyer NJ, Derzon JH, Winges L, et al. Effectiveness of practices to reduce blood sample hemolysis in EDs: a laboratory medicine best practices systematic review and meta-analysis. Clin Biochem. 2012;45(13-14):1012-1032.

(15.) Cox SR, Dages JH, Jarjoura D, Hazelett S. Blood samples drawn from IV catheters have less hemolysis when 5-mL (vs 10-mL) collection tubes are used. J Emerg Nurs. 2004;30(6):529-553.

(16.) Dwyer DG, Fry M, Sommerville A, Holdgate A. Randomized single blind control trial comparing haemolysis rate between two cannula aspiration techniques. Emerg Med Australas. 2006;18(5-6):484-488.

(17.) Munnix ICA, Schellart M, Gorissen C, Kleinveld HA. Factors reducing hemolysis rates in blood samples from the emergency department. Clin Chem Lab Med. 2011;49(1):196-197.

(18.) Kennedy C, Angermuller S, King R, et al. A comparison of hemolysis rates using intravenous catheters versus venipuncture tubes for obtaining blood samples. J Emerg Nurs. 1996;22(6):566-569.

(19.) Grant MS. The effect of blood drawing techniques and equipment on the hemolysis of ED laboratory blood samples. J Emerg Nurs. 2003;29 (2):116-121.

(20.) Ong ME, Chan YH, Limb H. Reducing blood sample hemolysis at a tertiary hospital emergency department. Am J Med. 2009;122(11):1054-el-6.

(21.) Proehl JA, Bradford JY, Leviner S, et al. Clinical practice guideline: prevention of blood specimen hemolysis in peripherally-collected venous specimens. Des Plaines, IL: Emergency Nurses Association. http://www.ena.org/ practiceresearch/research/CPG/Documents/HemolysisSynopsis.pdf. Published December 2012. Accessed July 15, 2014.

(22.) Hawkins RC. Poor knowledge and faulty thinking regarding hemolysis and potassium elevation. Clin Chem Lab Med. 2005;45(2):216-220.

(23.) Green SF. The cost of poor blood specimen quality and errors in preanalytical processes. Clin Biochem. 2013;46 (13-14):1175-1179.

Peter J. Howanitz, MD; Christopher M. Lehman, MD; Bruce A. Jones, MD; Frederick A. Meier, MD; Gary L. Horowitz, MD

Accepted for publication September 24, 2014.

From the Department of Pathology at SUNY Downstate, Brooklyn, New York (Dr Howanitz); the Department of Pathology, University of Utah, Salt Lake City (Dr Lehman); the Department of Pathology and Laboratory Medicine, Henry Ford Health System, Detroit, Michigan (Drs Jones and Meier); and the Department of Pathology, Beth Israel Deaconess Medical Center, Boston, Massachusetts (Dr Horowitz).

The authors have no relevant financial interest in the products or companies described in this article.

This manuscript was developed as a joint project by members of the Quality Practices and the Chemistry Resource Committees of the College of American Pathologists.

Reprints: Peter J. Howanitz, MD, Department of Pathology, SUNY Downstate, MS25, 450 Clarkson Ave, Brooklyn, NY 11203 (e-mail: Peter.Howanitz@downstate.edu).

Caption: Figure 6. Percentage of participants using the hemolysis scale for lipemia and icterus for lactate dehydrogenase (LD; n =627), glucose (n = 662), and potassium (n =684).

Please Note: Illustration(s) are not available due to copyright restrictions.
Figure 1. Specimen rejection and standardization
policies when specimens were hemolyzed
for lactate dehydrogenase (LD),
glucose, and potassium (n =846).

                       Laboratories, %

             Written policy of     Standardized reports
             specimen rejection    between primary and
                                    secondary analyzers

LD                 69                      40
Glucose            55                      40
Potassium          85                      46

Note: Table made from bar graph.

Figure 2. Percentage of participants who attempted validation of
manufacturers' determination of the effects of hemolysis for
lactate dehydrogenase (LD), glucose, potassium, or any analyte (n
=694).

LD          23
Glucose     24
Potassium   29
Any         30

Note: Table made from bar graph.

Figure 3. Percentage of clinical laboratories
taking action on hemolysis during past year (n
= 692).

Took corrective action past year       49
to decrease hemolysis

Systematically regularly monitored     47
hemolysis

Systematically regularly monitored     13
hemolysis specific locations

Systematically regularly monitored     34
hemolysis all locations

Note: Table made from bar graph.

Figure 4. Most common corrective interdepartmental
actions taken during the past year
to decrease hemolysis (n =337 laboratories,
806 actions).

Other                                            18

Provided information for physicians              19

Established quality improvement team between     37
laboratory and problem locations

Provided ongoing training for phlebotomists      53

Troubleshot outliers and discussed hemolysis     55
avoidance with phlebotomists

Collected hemolysis data and distributed to      57
leadership where hemolysis has occurred

Note: Table made from bar graph.

Figure 5. Participants' lack of success in
reducing hemolysis (N =717).

Other                                            30
Now ignore hemolysis                             0.3
Given up on reducing because nothing helps       2
Continue lo try to reduce, bul nothing helps     8
Concerned about percentage but do not know how   10
  to reduce
Try to reduce and have made slow progress        50

Note: Table made from bar graph.
COPYRIGHT 2015 College of American Pathologists
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2015 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Howanitz, Peter J.; Lehman, Christopher M.; Jones, Bruce A.; Meier, Frederick A.; Horowitz, Gary L.
Publication:Archives of Pathology & Laboratory Medicine
Article Type:Survey
Date:Jul 1, 2015
Words:4226
Previous Article:Pathophysiology and Diseases of the Proximal Pathways of the Biliary System.
Next Article:Design of a Genomics Curriculum: Competencies for Practicing Pathologists.
Topics:

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters