From the Department of Ophthalmology, University of California, SanFrancisco (Drs Kniestedt, Nee, and Stamper); and Department of Ophthalmology,University Hospital Zurich, Zurich, Switzerland (Dr Kniestedt). The authorshave no relevant financial interest in this article.
To compare intraocular pressure measurements obtained by recently introduceddynamic contour tonometry (DCT), Goldmann applanation tonometry (GAT), pneumatonometry(PTG), and intracameral manometry in human cadaver eyes.
Sixteen freshly enucleated human cadaver eyes were deepithelializedand dehydrated with dextran. A tube was placed in the anterior chamber andconnected to a transducer and to a bottle system filled with balanced saltsolution. The pressure in the eye was then altered between 5 mm Hg and 58mm Hg. Intraocular pressure measurements were obtained with DCT, GAT, andPTG at each manometric pressure reading.
On average, DCT values measured 0.58 ± 0.70 mm Hg higher thanreal intracameral pressure. The GAT and PGT showed consistently lower values,−4.01 ± 1.76 mm Hg and −5.09 ± 2.61 mm Hg, respectively.At all bottle heights, DCT values were significantly closer to the referencepressure than GAT and PTG (P<.001).
Measurement with DCT provides IOP values significantly closer to truemanometric levels than either GAT or PTG. Further studies are warranted todetermine its reliability in patients and the effect of corneal thickness.
Goldmann applanation tonometry (GAT) has become the gold standard againstwhich other tonometers have been compared.1- 7 Thisis despite the fact that Goldmann himself recognized that the tonometer'saccuracy was questionable in corneas that were not of average thickness.1 Using optical pachymetry, Goldmann and Schmidt1 assumed that most corneas were about 520 µmthick. Recent evidence has confirmed the wide variation in corneal thicknessand raised the possibility that corneas that are thicker or thinner than averagemight be associated with clinically significant overestimates or underestimatesof intraocular pressure (IOP) when measured with the GAT.8- 19 Somestudies suggested that at least some cases of normal-tension glaucoma couldbe explained by the underestimation of IOP by GAT in these patients whosecorneas are thinner than those with open-angle glaucoma.13,20- 22
Other tonometers, such as the pneumatonometer (PTG),23,24 theMcKay-Marg tonometer, and its derivative, the handheld tonometer (Tono-Pen;Mentor, Norwell, Mass), have been reported to be more accurate, especiallyin corneas outside the average in thickness.12,25 However,these observations have not always been confirmed by others.26- 30 Recently,a new device, the dynamic contour tonometer (DCT), was introduced by Kanngiesserand Robert.31 Dynamic contour tonometry isa totally new method of measuring the IOP based on the principle of contourmatching. The DCT tip is mounted in a tip holder of a GAT (Figure 1A), which provides for a constant appositional force of1 g. The contoured tip features a concave surface with a radius of 10.5 mm,a contact surface of approximately 7 mm, and a miniaturized piezoresistivepressure sensor 1.7 mm in diameter built flush into the center of the contactsurface (Figure 1B). Forcing thecentral area of the cornea into the contour of the DCT tip allows the examinerto measure the pressure of the eye directly and continuously on the externalsurface of the cornea. Pressure readings are sampled and digitized at 100Hz. Data points are transferred to a microprocessor-based control unit, whichcomputes and displays the measured pressure.
Dynamic contour tonometer (DCT)tip. A, The DCT tip inserted into a Goldmann applanation tonometer tip holder.B, Surface of the experimental DCT tip. A piezoelectric pressure sensor (1)is built in the concavity of the tip. The contour-matched contact surface(2) is located concentrically around the pressure sensor.
The purpose of this study was to compare the DCT, GAT, and PTG withmanometric measurements in human cadaver eyes.
Twenty freshly enucleated human cadaver globes that were unsuitablefor transplantation were provided by the eye banks of San Diego, Calif, andPortland, Ore. Eyes were stored at 4°C in moist chambers until use. Consentfor the use of the eyes was obtained by the eye banks either through premorteminstructions or from postmortem next of kin. Institutional review board consentwas not considered necessary for this in vitro study, but departmental approvalwas obtained. The eyes were used for this experiment on the day of arrival,an average ± SD of 27.0 ± 10.9 hours post mortem. The ages ofthe cadavers ranged from 64 to 84 years. The causes of death were either cardiovascularor neoplastic (other than ocular). Eyes with a history or evidence of previousintraocular surgery or corneal abnormalities were excluded. Four eyes servedfor a preliminary study evaluating closed vs open stopcock systems, and theremaining 16 eyes were included in the main study and final analysis.
Central corneal thickness (CCT) was measured with an ultrasonic pachymeter(Model 850; Humphrey Instruments, San Leandro, Calif), which was calibratedbefore each use. The speed of sound was adjusted at 1640 m/s according tothe internationally accepted standard velocity for human corneas and did notrequire adjustment for corneal thickness differences. Five independent measurementsin micrometers, their mean and standard deviation, and the 95% confidenceinterval (CI) were recorded.
All 20 eyes were subject to corneal dehydration, since the untreatedcorneal thicknesses varied between 444 and 925 µm depending, in part,on the postmortem age of the eye. The globe was first brought to an IOP of25 to 30 mm Hg by carefully injecting a 20% dextran solution into the anteriorchamber by means of a scleral entry with a 30-gauge needle. The 20% dextranwas prepared by dissolving 20 g of 229 000–molecular weight dextran(Sigma-Aldrich Corp, St Louis, Mo) in 100 mL of balanced salt solution (Alcon,Ft Worth, Tex). The corneal epithelium was totally removed by manual abrasionwith a cotton swab and a blunt scalpel blade. Corneal thickness was measuredfor the first time immediately after dextran injection. Next the globe wasimmersed in dextran solution. The ultrasonic CCT was measured after 15, 30,and 45 minutes of immersion. After 45 minutes, the ultrasonic CCT became relativelystable, at which time the anterior and posterior corneal curvature, astigmatism,and (optical) CCT were determined with a corneal topography system (OrbscanII; Bausch & Lomb Surgical, St Louis).
A 22-gauge needle with Y-adaptor (Saf-T-Intima, Vialon; Becton, Dickinsonand Company, Franklin Lakes, NJ) was then inserted into the anterior chambervia a separate scleral approach. Extreme care was taken with all penetrationsof the eye to avoid touching the endothelium, the iris, or the lens. The entireglobe was mounted in an eye holder embedded in moisturized gauze facing aslitlamp (30 SL-M; Carl Zeiss Meditec AG, Jena, Germany). The tube was connectedto a manometric transducer, an isotonic sodium chloride solution infusionbottle, and an open-air reference tube. Multiple stopcocks were also attachedto bleed all bubbles from the system and to allow either open or closed stopcocktechniques. The transducer and the anterior chamber were kept at the sameheight. The isotonic sodium chloride solution infusion bottle was attachedto an electrically driven intravenous pole for bottle height adjustment (Figure 2). The same instrument (Model 30Classic; Medtronic Inc, Minneapolis, Minn) was used for all PTG readings.
Experimental tubing system. 1,Dynamic contour tonometer (DCT) base station; 2, DCT recharge unit; 3, DCTpressure-sensitive tip; 4, Goldmann applanation tonometry tip holder mountedat a slitlamp; 5, human cadaver eye, anterior chamber filled with 20% dextran;6, eye holder with moisturized gauze; 7, pressure transducer; 8, manometricdevice; 9-11, tubing system filled with balanced salt solution; 9, referencetube open to atmospheric pressure; 10, bottle with isotonic sodium chloridesolution, variable height; 11, stopcocks with syringes to bleed air bubblesfrom the tubing system. CCW 3-6 indicates that items labeled 3, 4, 5, and6 were rotated 90° during the actual measurements of intraocular pressurewith all of the instruments but are shown in the displayed orientation forsimplicity of diagram.
In the preliminary series of 4 eyes, the pressure measurements wereperformed by means of the closed stopcock tubing system first to compare itsapplicability and accuracy with the open tubing system, which was used onthe same eye immediately thereafter. The bottle height was altered between5 cm and 50 cm and pressure readings were taken with DCT, PTG, and a handheldtonometer (Tono-Pen XL; Mentor) (TPN).
Sixteen human cadaver eyes were used for the main study. The bottleheight was adjusted at 5 cm to begin each trial. At each bottle height, 5independent measurements of DCT, GAT, and PTG were made in randomized orderwith manometric pressure measured and recorded between each set of measurements.The bottle heights were increased in 5-cm increments up to a maximum of 75cm (58 mm Hg). After the experiment, the bottle height was lowered to theinitial 5 cm. The series was accepted only if the initial and closing manometricpressures were within ±1 mm Hg.
To determine whether elevated IOPs alter the corneal curvature and theaccuracy of DCT, 3 eyes were subjected to optical keratometry and cornealtopography (Orbscan II) at a maximal IOP of approximately 58 mm Hg after thelast trial series. The stopcock at the 22-gauge needle tube was closed atthe maximal IOP and the globe was left in the cadaver eye holder but unhookedfrom the reservoir bottle distal to the stopcock. The cadaver eye holder wasmounted to the topography instrument and fixed in a way similar to that doneat the slitlamp for the IOP measurement trials. An average of 2 scans of cornealthickness (measured in micrometers) and anterior and posterior curvature (inmillimeters and diopters) were taken of each cornea, ensuring that the corneawas well moisturized and the entry needle without torsion during the readings.
Statistical analysis was performed with a homoscedastic, 2-tailed t test. The t tests were run atall bottle heights separately and all 5 pressure readings were taken intocalculation (n = 1120 for GAT and PTG, bottle heights of 10-75 cm; n = 1200for DCT, bottle heights of 5-75 cm). A mixed-effects regression model wasconstructed by means of SAS software (SAS Institute Inc, Cary, NC). The modeltreated cadavers and eyes nested within cadavers as random effects and didnot assume equal variability in the 3 devices. A separate analysis was runusing only the data from trials with bottle heights from 10 to 30 cm (8-25mm Hg) to derive the statistics for the "clinically significant" range. Dataare given as mean ± SD unless otherwise indicated.
Four eyes were used for the comparison of the open vs closed tubingsystems. These eyes were not included in the analysis of the main study. Theopen system showed an overestimation by DCT (+0.54 ± 0.29 mm Hg) andan underestimation by PTG (−2.73 ± 1.94 mm Hg) and TPN (−4.92± 1.91 mm Hg). In a closed system, the cadaver globe is subject touncontrollable fluid leakage.23,24,32 Anykind of tonometry that flattens or indents the cornea displaces a variableamount of fluid within the anterior chamber or to other compartments of theglobe, which would result in a transient increase in IOP followed by a decreasedue to a consequent further loss of fluid from the globe through the normaloutflow pathways (tonographic effect) as well as through cut ends of the venoussystem. Therefore, it is difficult to determine at what point to measure themanometric IOP: either before probe application or during applanation (byPTG or TPN) or while contour matching (DCT) is taking place. The DCT valuesmeasured −0.11 ± 0.72 mm Hg; PTG, −2.92 ± 1.08 mmHg; and TPN, −3.98 ± 2.53 mm Hg compared with the starting pressure(before probe application). With the use of continuous manometric pressuremeasurement, the results were −0.59 ± 0.31 for DCT, +3.01 ±1.13 for PTG, and −4.53 ± 2.10 for TPN. Regardless of which pointin time the tonometric pressure was compared with manometry using the closedstopcock method, DCT was significantly closer to manometric reference pressurethan either PTG or TPN (P<.001). Furthermore,the data show little difference in results between open and closed stopcockmethods.
The 16 dehydrated and untreated human cadaver corneas presented initiallywith a CCT of 805 ± 102 µm (95% CI, 755-855 µm). After15, 30, and 45 minutes of exposure of both sides of the cornea to 20% dextran,CCT of 617 ± 85 µm (95% CI, 575-659 µm), 512 ± 59µm (95% CI, 483-541 µm), and 462 ± 49 µm (95% CI,438-486 µm) was achieved. At the end of the experiment, CCT showed furtherthinning to 434 ± 48 µm (95% CI, 411-457 µm). The medianCCT was fairly stable until the IOP of 50 mm Hg was reached, falling backslightly in the markedly elevated pressure range.
Mean examination duration after dehydration was 85 ± 19 minutes.Anterior corneal curvature was found to be 7.98 ± 0.46 mm (42.47 ±1.79 diopters); posterior curvature was 6.40 ± 0.40 mm. The CCT determinedby the corneal topography system was 489 ± 103 µm—surprisingly,27 µm thicker than the CCT determined by ultrasonic pachymetry. Thediscrepancy may be due to the larger scatter from the topography device andthe difficulty in centering the cadaver globe in the front of the laser scanningarea.
Three eyes were subjected to CCT measurement by the orbital topographysystem when the IOP was greater than 55 mm Hg. In these eyes, anterior cornealradius before the experiment at an IOP of approximately 20 mm Hg was 7.71± 0.21 mm (43.17 ± 1.33 diopters) and posterior curvature measured5.85 ± 0.04 mm (57.4 ± 0.3 diopters). At the end of the experiment,at an IOP of about 58 mm Hg, the anterior curvature was 7.69 ± 0.18mm (43.2 ± 1.3 diopters) and the posterior curvature, 6.4 ±0.21 mm (52.8 ± 1.7 diopters). For the anterior curvature, this changewas not statistically significant (P = .31), butthe posterior curvature flattened at the borderline of statistical significance(P = .03).
At each bottle height, all 3 tonometer devices were used to obtain 5pressure readings. The bottle height was altered in 5-cm increments between5 cm and 75 cm, except for GAT and PTG. Neither GAT nor PTG was able to produceconsistent readings at the 5-cm bottle height. The IOP readings (with standarddeviations) at defined bottle heights for manometry, DCT, GAT, and PTG arelisted in Table 1. Compared withthe absolute manometric pressure, DCT measures were not significantly different(P = .18, n = 1200). The GAT and PTG measures wereboth significantly different from the manometric pressure (P<.001, n = 1120), as shown in Figure 3.
Absolute pressure readings. Intraocularpressure (IOP) values obtained with the dynamic contour tonometer (DCT) didnot differ significantly from the manometric pressure (P = .18). Goldmann applanation tonometry (GAT) and pneumatonometry(PTG) provided values significantly different from the manometric referencepressure (IOP) (P<.001).
With manometric pressure used as the reference, DCT showed a slightabsolute deviation between 0.3 mm Hg (95% CI, 0.2-0.4 mm Hg) at 20-cm bottleheight (16.2 mm Hg) and 1.7 mm Hg (95% CI, 1.1-2.3 mm Hg) at 75-cm bottleheight (58 mm Hg), whereas GAT underestimated the IOP with an absolute deviationbetween −3.2 mm Hg (95% CI, 2.6-3.8 mm Hg) at 20-cm bottle height (16.2mm Hg) and −5.9 mm Hg (95% CI, 4.7-7.1 mm Hg) at 70-cm bottle height(58 mm Hg). The PTG was rather accurate in the low end of the pressure spectrum(absolute deviation, 1.1 mm Hg; 95% CI, 0.6-1.6 mm Hg) at 10-cm bottle height(8.5 mm Hg), but fell off with a significant underestimation by 9.6 mm Hg(95% CI, 8.8-10.4 mm Hg) at 75-cm bottle height (58 mm Hg). On average, DCTmeasured +0.58 ± 0.70 mm Hg higher than the manometric pressure. TheGAT and PTG showed consistently lower average values: −4.01 ±1.76 mm Hg and −5.09 ± 2.61 mm Hg, respectively (Table 1 and Figure 4;algebraic accuracy).
Algebraic deviation from manometricintraocular pressure (IOP). Values obtained with the dynamic contour tonometer(DCT) measured on average 0.58 ± 0.70 mm Hg higher than intracameralpressure indicated by the manometer. Goldmann applanation tonometry (GAT)and pneumatonometry (PGT) showed consistently lower values, −4.01 ±1.76 mm Hg and −5.09 ± 2.61 mm Hg, respectively. Dynamic contourtonometry (DCT) values at all bottle heights were significantly closer tothe reference pressure than GAT and PTG values (P<.001).
A precise assessment of the IOP is usually most critical in clinicalsituations in which the pressure range is between 10 and 25 mm Hg. Withinthis range, DCT showed an average of +0.33 ± 0.49 mm Hg; GAT, −3.43± 1.24 mm Hg; and PTG, −2.97 ± 1.82 mm Hg compared withmanometric readings (algebraic accuracy). Although all 3 tonometers performedreasonably well in this range, the DCT was considerably more accurate. Infact, DCT was significantly closer to manometric pressure throughout all bottleheights than GAT and PTG, with P values always lessthan .001. The GAT and PTG readings had not only a larger IOP error than DCTbut also greater variability.
Scattergrams and linear regression were plotted for DCT, GAT, and PTGfor pressure values in the ranges 8 to 28, 8 to 38, and 8 to 58 mm Hg, respectively,in the hope of finding a linear correlation between corneal thickness andIOP values. No significant correlation could be found between the real IOPand the degree of error of any of the tonometers. The strongest correlationwas obtained in the range of 8 to 28 mm Hg with slopes of 0.001 for DCT (r = 0.089), −0.003 for GAT (r =0.111), and −0.0014 for PTG (r = 0.036). Giventhe relatively narrow range of CCT measurements due to artificially distributedcorneal thickness, significant correlation would be difficult to obtain.
The results of this study strongly suggest that DCT is superior in accuracyin eye bank eyes to GAT and PTG across the range of IOPs likely to be foundin clinical practice. Furthermore, the accuracy seems to be independent ofcorneal thickness and corneal curvature (within the limits of this study).In the low- and moderate-pressure range, inaccuracy of the GAT and the PTGwas not great and not likely to be of major clinical significance, but alsowas not linear. However, in the range of 18 to 23 mm Hg, an underestimationof 3 or 4 mm Hg could be clinically problematic, and the increasing underestimationof IOP by PTG in the high-pressure range (35-58 mm Hg) might also be clinicallysignificant.
Goldmann and Schmidt1- 3 themselvespointed out that applanation tonometry would be affected by corneal rigidityand thickness and set the applanation diameter at 3.06 mm so that the effectof corneal elasticity in an "average" cornea would offset the effect of thecapillary attraction of tears. On the basis of relatively small numbers ofpatients and optical pachymetry, they concluded that "most" patients had aCCT between 500 and 520 µm.1,2 Ultrasonicpachymeters tend to give greater corneal thickness measurements than opticalpachymeters, perhaps because the latter "ignore" the epithelium and endothelium.33 Recent studies have shown that there is a variationof almost 200 µm in corneal thickness.8,13- 16,22,34 Itis likely that significant errors may be introduced by applanation tonometersin corneas significantly different from the "average" that Goldmann and Schmidtassumed.
Performing applanation tonometry with the GAT on cadaver eyes is challengingand fraught with potential for error. Corneal epithelial abnormalities aresome of the most common sources of error in applanation tonometry.35 Because corneal epithelium was susceptible to damageduring our initial preliminary study, we decided to remove corneal epitheliumthoroughly in all eyes used for the study.36,37 Theeffect of absence of corneal epithelium on the accuracy of GAT has not beeninvestigated and is still unknown.10 To ourknowledge, Goldmann did not describe how to adjust the mires for GAT performedon the Bowman layer directly. On freshly enucleated cadaver eyes with cornealthicknesses of 475 to 500 µm, on which GAT has been shown to be mostaccurate (keeping in mind that approximately 50 µm of corneal epitheliumis removed), we found the best correlation to the manometric intracameralpressure by using the semicircle adjustment seen in Figure 5. In the absence of corneal epithelium, the usual fluoresceinhemirings cannot be detected. We used full semicircles that were brought intoan overlapping position of approximately twice the width that Goldmann andSchmidt proposed for the ring.2,4
Goldmann applanation tonometryhemidisc setting chosen for the entire study. Hemidiscs overlap approximatelyone fifth of the disc diameter (2 times the ring width).
Controversy exists regarding the method to manometrically measure IOPin cadaver eyes. Proponents of the closed stopcock technique claim greateraccuracy.24,32 However, in anenucleated eye, any displacement of fluid such as occurs with applanationtonometry or PTG will first cause an elevation of IOP and then a slow decline(similar to tonography) due to leakage of fluid from the normal aqueous outflowpathways as well as the cut ends of choroidal and episcleral vessels. Findingthe right point in this changing environment at which to determine the "true"manometric pressure is difficult if not impossible. Most studies have usedaverage pressures or some kind of algorithm to peg the manometric pressure.23,24,32,38
In this study, we tried to address the issue of an open vs a closedstopcock system by using both methods in a preliminary study performed on4 human cadaver eyes defined by the same inclusion criteria as in the mainstudy. In the closed system, the stopcock is closed immediately before thepressure reading by whatever device to simulate in vivo conditions and toreduce the effect of pressure-dependent fluid leakage out of the eye (approximately5 mm Hg/min at an IOP of 45 mm Hg, and 2 mm Hg/min at 20 mm Hg). Pneumatonometryespecially causes the initial rise in pressure by as much as 10 mm Hg. Itis problematic which manometric pressure to pick and compare with the IOPobtained by PTG and TPN. On average, both PTG and TPN underestimated IOP comparedwith the actual manometric IOP peak (−2.29 mm Hg and −3.98 mmHg, respectively). The DCT showed no significant difference in accuracy withthe closed and open stopcock system, probably because the DCT displaces anegligible amount of fluid from the anterior chamber. The similar resultsobtained with both open and closed stopcock systems suggest that our observationsare valid.
In this cadaver eye model, DCT is more likely to be consistent withthe manometric pressure than both GAT and PTG. The use of cadaver eyes asthe model did not appear to affect the accuracy of DCT. The use of DCT wasquick and easy. Since a value is given automatically by the strain gauge andthe automatic software in the base station, there is less chance for observerbias with this system than with GAT. Similarly, PTG gives a more objectivevalue.
The use of a cadaver model introduces some potential error, but, notingthe very small P values (<.001), DCT seems toaccurately reflect the internal pressure. The data also suggest that DCT isunaffected by the loss of the corneal epithelium in this model. The DCT doesseem to measure slightly higher than manometric pressure at markedly elevatedIOP (>60 mm Hg). It is possible that this slight inaccuracy may be due tochanges in the corneal curvature as the pressure becomes very high. However,in the 3 eyes in which orbital topographic measurements were obtained, nosuch changes occurred in the anterior curvature. This was a very small sampleand might not be representative of the whole group, or other unknown histopathologicchanges in the anterior part of the cornea may have occurred that were notdetectable by the orbital topography system in these stressed eyes. The devicedid detect a significant flattening in the posterior curvature of the corneawhen the IOP was extremely high.
In eyes with very elevated IOP, both GAT and PTG measured consistentlytoo low compared with manometry. Applanation tonometry is certainly knownto be influenced by corneal thickness. The eyes used in this study reacheda CCT of 462 µm after dehydration; this is approximately 60 to 80 µmthinner than in living conditions. Furthermore, the corneas are subject tothinning even more during the experiments because of the dextran environmentand evaporation, perhaps explaining the consistently low GAT and PTG values.In this sample of cadaver eyes, regression lines showed very small slopesand no significant linear correlation between CCT and IOP as measured by anydevice. This suggests that even in very thin corneas, DCT maintains a satisfactoryrepresentation of true IOP.
These results strongly suggest that DCT is a promising technology thatmay afford more accurate IOP measurement across the range of IOPs found inclinical practice. Further studies are indicated to further elucidate theproperties, usefulness, and accuracy of DCT in clinical situations.
Correspondence: Robert L. Stamper, MD, Department of Ophthalmology,University of California, San Francisco, 10 Kirkham St, PO Box 730, San Francisco,CA 94143 (email@example.com).
Submitted for publication August 11, 2003; final revision received March31, 2004; accepted April 6, 2004.
This study was supported by the Herman-Klaus Fund, Zurich, Switzerland,and That Man May See Foundation, San Francisco, Calif. Dr Kniestedt is financiallysupported by the Swiss National Fond, Bern, Switzerland, and the Swiss Fundto Prevent and Fight Against Blindness, Zurich.
We thank Swiss Microtechnology AG, Port, Switzerland, for providingthe experimental dynamic contour tonometer and Alan Bostrom, PhD, from theDepartment of Biostatistics at University of California, San Francisco, forthe statistical analysis. We also thank Robin Troyer Olson for her supportin the laboratory and for her coordination with the eye bank centers.
Thank you for submitting a comment on this article. It will be reviewed by JAMA Ophthalmology editors. You will be notified when your comment has been published. Comments should not exceed 500 words of text and 10 references.
Do not submit personal medical questions or information that could identify a specific patient, questions about a particular case, or general inquiries to an author. Only content that has not been published, posted, or submitted elsewhere should be submitted. By submitting this Comment, you and any coauthors transfer copyright to the journal if your Comment is posted.
* = Required Field
Disclosure of Any Conflicts of Interest*
Indicate all relevant conflicts of interest of each author below, including all relevant financial interests, activities, and relationships within the past 3 years including, but not limited to, employment, affiliation, grants or funding, consultancies, honoraria or payment, speakers’ bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued. If all authors have none, check "No potential conflicts or relevant financial interests" in the box below. Please also indicate any funding received in support of this work. The information will be posted with your response.
Register and get free email Table of Contents alerts, saved searches, PowerPoint downloads, CME quizzes, and more
Subscribe for full-text access to content from 1998 forward and a host of useful features
Activate your current subscription (AMA members and current subscribers)
Purchase Online Access to this article for 24 hours
Some tools below are only available to our subscribers or users with an online account.
Download citation file:
Web of Science® Times Cited: 84
Customize your page view by dragging & repositioning the boxes below.
and access these and other features:
Enter your username and email address. We'll send you a link to reset your password.
Enter your username and email address. We'll send instructions on how to reset your password to the email address we have on record.
Athens and Shibboleth are access management services that provide single sign-on to protected resources. They replace the multiple user names and passwords necessary to access subscription-based content with a single user name and password that can be entered once per session. It operates independently of a user's location or IP address. If your institution uses Athens or Shibboleth authentication, please contact your site administrator to receive your user name and password.