Diagnostic accuracy of secretin-stimulated ultrasonography of the pancreas assessing exocrine pancreatic failure in cystic fibrosis and chronic pancreatitis

Scandinavian Journal of Gastroenterology. 2015; Early Online, 1–10

ORIGINAL ARTICLE

Diagnostic accuracy of secretin-stimulated ultrasonography of thepancreas assessing exocrine pancreatic failure in cystic fibrosis andchronic pancreatitis

TROND ENGJOM1,2, FRIEDEMANN ERCHINGER2,3, ERLING TJORA2,4,BIRGER N. LÆRUM2,5, DIMCEVSKI GEORG1,2 & ODD H. GILJA2,6

1Department of Medicine, Haukeland University Hospital, Haukeland University Hospital, 5021 Bergen, Norway,2Department of Clinical Science, University of Bergen, Bergen, Norway, 3Department of Medicine, Voss Hospital,Sjukehusvegen 16, 5700 Voss, Norway, 4Pediatric Department, Haukeland University Hospital, 5021 Bergen, Norway,5Department of Thoracic Medicine, Haukeland University Hospital, 5021 Bergen, Norway, and 6National Centre forUltrasound in Gastroenterology, Haukeland University Hospital, 5021 Bergen, Norway

AbstractObjective. Volume output failure is a feature of decreasing exocrine pancreatic function. This parameter is assessed bysecretin-stimulated MRI in several studies. Our purpose was to evaluate ultrasonography of the fluid in the descendingduodenum andWirsung duct (WD) after secretin stimulation as a measure of pancreatic fluid flow in patients expected to havesevere output failure.Material andmethods.We included subjects with chronic pancreatitis (CP), cystic fibrosis (CF) and agroup of healthy controls in a prospective observation study. Transabdominal ultrasonography was performed before andduring 15 min after secretin i.v. duodenal juice was collected by endoscopic short test (EST), and bicarbonate concentrationmeasured. Patient groups were classified according to exocrine pancreatic function. Results. Pancreatic insufficient CF (CFI)patients and CP insufficient (CPI) patients showed less duodenal fluid filling compared to other groups (p < 0.001). Measuresof the WD diameter could only identify the most severe failure in the CFI group (p < 0.001). Conclusion. Secretin-stimulatedultrasonography can be used to assess pancreatic fluid flow and may be combined with EST in the evaluation of exocrinepancreatic function. Fluid filling in the descending part of duodenum was the most accurate predictor of pancreaticinsufficiency in both patient groups. The test demonstrated better diagnostic accuracy diagnosing exocrine pancreatic failurein the CF patients than in CP patients.

Key Words: chronic pancreatitis, cystic fibrosis, exocrine pancreatic function, gastrointestinal ultrasonography

Introduction

The literature on exocrine-pancreatic function testinghas focused little on differences in ductal and acinarcausesofpancreatic failure.Following theemergenceofnew treatments using modulators of the cystic fibrosis(CF) transmembrane receptor protein (CFTR) inspecific genetic subtypes of the CF population, thereis a need for simple clinical tools capable of assessingductal or CFTR function both in adults and children[1]. Parameters of pancreatic CFTR function are

pancreaticoutputvolumeandpancreatic-juicecontentsof chloride andbicarbonate [2]. Pancreaticfluidflow inresponse to secretin stimulation in humans has beenstudied by various collecting methods of pancreaticsecretions, secretin-stimulated magnetic resonancecholangio-pancreatography (MRCP) (s-MRCP) andbydirect intubationof thepancreaticductduring endo-scopic retrograde pancreatography (ERCP) [3-7].Dynamics, volume and content of the secretions arewell known and related to grades of exocrine failure andvarious pathological conditions [6,8-13].

Correspondence: Trond Engjom, Department of Clinical Medicine, University of Bergen, Haukeland University Hospital, 5021 Bergen, Norway.Tel: +47 92048 9852. Fax: +47 5597 2950. E-mail: [email protected]

(Received 6 June 2014; revised 4 November 2014; accepted 8 November 2014)

ISSN 0036-5521 print/ISSN 1502-7708 online � 2015 Informa HealthcareDOI: 10.3109/00365521.2015.1004363

Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y H

else

Ber

gen

- H

auke

land

uni

vers

itets

syke

hus

on 0

1/26

/15

For

pers

onal

use

onl

y.

Direct, quantitative tests collecting duodenal juiceafter stimulation of the pancreas with cholecystokinin,secretin or standardized meal and analyzing duodenalfluid for bicarbonate and/or digestive enzymes arethe most sensitive methods to identify early stages ofexocrine pancreatic failure (EPF) [11,14]. Endoscopicshort test (EST) measuring timed peak concentrationsof bicarbonate/digestive enzymes simplify the proce-dures and minimize patient discomfort, and are beingevaluated with promising results [15-19]. ESTs givelimited information on volume in pancreatic secretion.s-MRCP, imaging fluid filling of the pancreatic

duct and visual or mathematical models estimatingtotal duodenal fluid filling give an estimate of pan-creatic and duodenal fluid transport [3,5,7]. How-ever, limited accesses to MR time combined withtechnical complexity of the test are factors limitingthe distribution of s-MRCP.Ultrasonography offers high temporal and spatial

resolution and real-time imaging of the pancreas andsurrounding tissue, and can easily be repeated safelymany times [20]. Earlier studies have demonstratedthat changes in caliber of the Wirsung duct (WD)could be visualized by transabdominal ultrasonogra-phy following secretin stimulation, and some wereable to differentiate patients with chronic pancreatitis(CP) from healthy volunteers [21-23]. To our knowl-edge, the fluid filling of the descending duodenum(DD) assessed by ultrasonography in the diagnosis ofEPF has not been evaluated, and the diagnosticaccuracy of the secretin-stimulated ultrasonographyhas not been properly established.In this study, we used ultrasonography to assess

pancreatic fluid filling of the WD and DD aftersecretin stimulation. We aimed to evaluate the diag-nostic value of this test compared to exocrine pan-creatic function diagnosed by EST and fecal elastase(FE) in patient groups where EPF with reducedvolume output could be expected.

Methods

Subjects

During the period May 2009 to May 2014, patientsaged between 16 and 75 years with suspected CPreferred to our outpatient clinic were offered partic-ipation in a multimodal diagnostic protocol for thisexperimental study. The patients were scored accord-ing to a modified Mayo score described elsewhere[15,24]. We also included adult CF patients. The CFdiagnosis was evaluated according to the diagnosticcriteria for CF defined in the CF foundation consen-sus report [25]. A control group of healthy volunteerswas recruited by advertising through board notes intwo participating hospitals. Patients not able to giveinformed consent, and patients who had performedextensive pancreatic surgery, were not included. Anysubjects with insufficient ultrasonography visualiza-tion of the pancreas were excluded.The patient inclusion flow chart is illustrated in

(Figure 1). Fifty-one CP patients and 40 CF patientswere eligible and examined after informed consent.Twelve CP patients and three CF patients wereexcluded due to poor visualization of the pancreas.Twenty-seven healthy volunteers were also examined.One control person was unable to perform endoscopy,and was excluded. Accordingly, we included 39 CPpatients, 37CFpatients and 26 healthy volunteers.Weclassified the patients as exocrine pancreatic sufficientor insufficient by bicarbonate concentration in duo-denal juice and by FE-1 concentration. Patients withpeak bicarbonate concentration >80meq/l or f-elastase>100mg/g were considered pancreas sufficient. Whensorted by exocrine pancreatic function, patient groupswere divided as follows: chronic pancreatitis insuffi-cient (CPI): n = 15, chronic pancreatitis sufficient(CPS) n = 24, cystic fibrosis, pancreatic insufficient(CFI): n = 16, pancreatic sufficient (CFS): n = 21.

CF, n = 40

CF, n = 37 CP, n = 39

CP, n = 51 HC, n = 27

Exclusions (all groups)CF: n = 3, CP: n = 12, HC: n = 1(Poor visualisationor not complete protocol)

CFI, n = 16 CFS, n = 21 CPI, n = 15 CPS, n = 24 HC, n = 26

Figure 1. Flow chart displaying patient inclusion. Abbreviations: CF = Cystic fibrosis; CFI/CFS = Cystic fibrosis insufficient/sufficient; CP:Chronic pancreatitis; CPI/CPS = Chronic pancreatitis insufficient/sufficient; HC = Healthy controls.

2 Engjom Trond et al.

Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y H

else

Ber

gen

- H

auke

land

uni

vers

itets

syke

hus

on 0

1/26

/15

For

pers

onal

use

onl

y.

Relevant clinical data were recorded and supple-mental examinations were performed if necessary forclinical evaluation. Characteristics of the three groupsare displayed in (Table I).There were no significant differences between

patient groups and controls related to age, genderor body mass index. When dividing the patient pop-ulation into separate groups, the non-CF groups wereolder than the CF groups (p < 0.05). There was anincreased alcohol intake in healthy controls (HC)versus patient groups (p < 0.05). The number ofsmokers was significantly higher in CP group thanin CF patients (p < 0.05).

Methods

(Figure 2) illustrates an examination flow chart withthe combination of secretin-stimulated ultrasonogra-phy and EST.The subjects fasted overnight. Transabdominal

ultrasonography was performed in supine positionwith a transverse or oblique epigastric probe position.A GE Logic E9 scanner with a 1-5 MHz CRA-probewas used (GE medical systems and primary carediagnostics, Milwaukee, WI, USA). The defaultabdomen configuration was used to acquire theimages: Frequency 4 MHz, frame rate 15–22 f/s

300 1Minutes

US + secretin test Endoscopic secretin test

Inj. secretin Start gastroscopy

5 10 15 25 35

Collect duodenal juice

40 45

Figure 2. Examination flow chart. Ultrasound was performed during the first 15 min after secretin. Endoscopic aspiration of duodenal juicewas performed between 30 and 45 min after secretin stimulation.

Table I. Demographic data.

Patients Controls p

CFI (n = 16) CFS (n = 21) CPI (n = 15) CPS (n = 24) HC (n = 26)

Age (range)* 22 (16–52) 21 (16–70) 59 (37–74) 59 (21–79) 38(19–64) ns1

Gender (,/<) 7/9 10/11 9/6 12/12 11/15 nsBMI 21 ± 2.5 24 ± 3.1 22 ± 3.9 23 ± 4.2 24 ± 4.2 nsSmoke 0/16 0/21 10/15 4/24 4/26 <0.052

Alcohol* 0 (0–3) 1 (0–12) 0 (0–7) 0 (0–50) 2 (1–8) <0.05Peak HCO3- 14 ± 5.1 109 ± 25 33 ± 17 94 ± 23 110 ± 16 <0.0013

F-Elastase* 0 (0–17) 565 (169–733) 7 (0–57) 430 (26–777) 552 (299–942) <0.0013

Asp. vol 2.0 ± 1.5 8.2 ± 3.3 9.1 ± 3.5 8.1 ± 3.7 10.9 ± 3.1 <0.0014

Values in mean ± SD unless otherwise stated.Abbreviations: BMI = Body mass index; Asp. Vol = Volume aspirated from duodenum; ns = Not significant.1No difference between patients and controls. CF patients younger than CP patients.2More smokers in CP groups.3Higher alcohol intake in HC groups.4Difference between pancreas sufficient and insufficient groups.5Difference between CFI and all other groups.*Median (range).

Accuracy of secretin stimulated ultrasound 3

Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y H

else

Ber

gen

- H

auke

land

uni

vers

itets

syke

hus

on 0

1/26

/15

For

pers

onal

use

onl

y.

varying depth of scanning. A complete scanning of thepancreas was performed. WD was identified in thebody of pancreas ventral to the superior mesentericartery (Figure 3). In cases with severe caliber changesdue to CP, a section with normal (not dilated) caliberwas chosen and local land marks registered to secure astandardized point of measurement. The descendingpart of duodenum was scanned and an image dis-playing the largest possible fluid-filled area wasstored. Subsequently, the patient received secretin(Secrelux� 10 U/ml. Sanochemia DiagnosticsDeutschland GmbH, Neuss, Germany) intrave-nously, 1 CU/kg, with a maximum dose of 70 CUafter a test dose of 10 CU [26]. The ultrasonographymeasures were repeated and images stored after 1, 5,10 and 15 min. The longitudinal diameter of thegallbladder was measured before secretin, and after15 min. Two operators performed sonographic mea-surements. Images were stored digitally and laterreviewed. The maximal diameter of luminal fluidpockets was measured at level of the pancreatic

head. Area tracing of the fluid pocket in the DDwas taken from 13 patients in the CPI group, 19 inCPS group, 22 patients in the HC group and 36 CFpatients (Figure 4).Twenty-five minutes after secretin administration,

an EST was performed using an Olympus GIF160 scope (Olympus Europa GmbH, Hamburg, Ger-many). The method is described in detail elsewhere[19,20]. Peak value of bicarbonate concentration wasmeasured and used as estimate of exocrine pancreaticfunction. The aspirated volume during the 15 minsampling period was also registered. Operators didnot have knowledge of pancreatic function or detailedpatient information at time of examination.FE-1 was measured by a commercial monoclonal

analysis kit (ScheBo� Biotech, Giessen, Germany).FE1 < 100 mg/mg was considered pathological.The study was conducted in accordance to the

Helsinki II Declaration [27] and was approved bythe local ethics committee (registration numbers2010/2857-7).

LOGIQE9

LOGIQE9

SMA

A

DWD

SMA

Figure 3. The figure displays ultrasound of the pancreas with detail of theWirsung duct. Abbreviations: D = Fluid-filled duodenum; A =Aorta;SMA = Superior mesenterial artery; WD = Wirsung duct.

0 min

1 minLOGIQ

E9 LOGIQE9

LOGIQE9

LOGIQE9

10 min

15 min

Figure 4. Duodenal filling after secretin stimulation in pancreatic sufficient patient. Markings illustrate the duodenal diameter and theduodenal area tracing.

4 Engjom Trond et al.

Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y H

else

Ber

gen

- H

auke

land

uni

vers

itets

syke

hus

on 0

1/26

/15

For

pers

onal

use

onl

y.

Statistical analysis

All statistics where calculated in SigmaPlot 11,(Systat Software, Inc., San Jose, CA, USA).Normality was tested by Kolmogorov–Smirnovtest. The results are expressed as mean with standarddeviation if normally distributed or as median valueswith range if not. Simple comparisons betweengroupsweremade by student t-test orMann–WhitneyU-test as appropriate. Repeated multiple comparisonsof the data were done by paired two-way Anova/posthoc analysis with Bonferroni/Sidak corrections withfactors patient groups and time after secretin stimu-lation, and with Kruskal–Wallis analysis of varianceon ranks in non-normal distributions. Five percentlevel of statistical significance was used. In the eval-uation of diagnostic accuracy of the duodenal areatracing, we also calculated area under the individualcurves (AUC) for the individual patient curves ofduodenal area versus time after secretin. Acuracywas calculated using receiver operating curves(ROC) in Sigmaplot 11 [28].

Results

Visualization

Out of the total number of 118 eligible subjects,15 subjects were excluded due to poor visualizationof the pancreas. Additional 12 of the included subjectshad recordings unsuitable to perform the area scan-ning of the DD.The WD and duodenal measurements are dis-

played in (Table II) and (Figures 5 and 6).

Diameter of Wirsung duct

There was a significant increase in mean diameter ofthe WD during the first 5 min after secretin stimu-lation in the two pancreatic sufficient groups(p < 0.05). This was not seen in the pancreaticinsufficient groups. The CFI group showed lesssecretin-stimulated duct dilatation compared to HC(p < 0.05). We were not able to differ between the

Table II. Results.

Patients Controls p

CFI (n = 16) CFS (n = 21) CPI (n = 15) CPS (n = 24) HC (n = 26)

Measures on Wirsung duct (cm)WD0 0.12

(0.11–0.14)0.13

(0.11–0.15)0.12

(0.10–0.15)0.12

(0.10–0.17)0.14

(0.12–0.16) pCFI vs. all<0.05WD1 0.13

(0.12–0.14)0.16

(0.14–0.18)0.15

(0.12–0.19)0.14

(0.12–0.17)0.16

(0.13–0.23)WD5 0.14

(0.12–0.15)0.15

(0.14–0.17)0.15

(0.12–0.17)0.14

(0.12–0.17)0.17

(0.14–0.19)WD15 0.13

(0.11–0.14)0.13

(0.11–0.18)0.13

(0.11–0.15)0.13

(0.11–0.18)0.14

(0.12–0.17)Measures on descending duodenum (cm)DD0 0.32

(0.0–0.43)0.41

(0.0–0.76)0.0

(0–0.59)0.70

(0.37–0.87)0.0

(0–0.07)DD1 0.75

(0.56–1.12)0.94

(0.80–1.18)0.60

0–1.59)0.99

0.76–1.46)1.01

(0.64–1.37)DD5 0.82

(0.68–1.15)1.40

(1.12–1.66)1.40

(1.18–1.81)1.40

(1.14–1.95)1.49

(1.20–1.96)DD15 0.99

(0.71–1.49)1.77

(1.09–2.09)1.47

(1.15–2.02)1.61

(1.25–2.47)1.84

(1.69–2.17)Area of the descending duodenum(cm2)DA0 0.46

(0–0.70)0.44

(0–0.73)0.0

(0–0.35)0.84

(0.39–1.75)0.0

(0–0.35)pCFI/CPI vs. HC/CFS/CPS<0.05

Peak DA5–15 1.23(0.90–2.11)

4.23(3.20–7.42)

3.78(2.69–5.85)

4.94(3.32–6.78)

6.03(5.17–7.05)

AUCarea 16.3(6.97–22.1)

39.8(28.6–65.0)

39.7(19.6–52.9)

54.3(39.4–72.8)

57(46.4–65.7)

pCFI vs. all<0.05pCPI vs. HC=0.015

Table displaying results ultrasonographic measures: Values in mean (IQ range).Abbreviations: WDx = Wirsung duct diameter (cm) � minutes after secretin stimulation; DDx = Diameter of the descending duodenum(cm2)�minutes after secretin DAx = Area of the descending duodenum (cm2)�minutes after secretin stimulation; AUCarea = Area under thecurve derived from a curve tracing D-area versus time.

Accuracy of secretin stimulated ultrasound 5

Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y H

else

Ber

gen

- H

auke

land

uni

vers

itets

syke

hus

on 0

1/26

/15

For

pers

onal

use

onl

y.

other pancreas sufficient and insufficient groups usingthese measures.

Diameter of the descending duodenum

There was a significant increase in duodenal dia-meters both 10 and 15 min after secretin stimulationin all groups (p < 0.05). CFI patients had significantless dilatation of the duodenal diameter at 10 and

15 min after secretin compared to all other groups(p < 0.05). There were no significant differencesbetween the other groups.

Area scanning of the descending duodenum

By tracing the area of theDDand calculatingAUC,wewere able to make a good differentiation between boththe pancreatic insufficient groups (CFI and CPI) and

HC pHC vs CFI < 0.05

Ductal diameter after secretin stimulation

CFICFSCPICPS

0.20

0.18

0.16

0.14

0.12

0.10

Time after secretin stimulation (min)

Du

ctal

dia

met

er (

cm)

0 5 10 15

HC pAll vs CFI < 0.05

Duodenal diameter after secretin stimulation

CFICFSCPICPS

2.5

2.0

1.5

1.0

0.5

0.0

Time after secretin stimulation (min)

D-D

iam

eter

(cm

)

0 5 10 15

Figure 5. Figures displaying diameters of the pancreatic duct (left) and descending duodenum (right) before and in defined time intervals aftersecretin stimulation (median values). Thesemeasurements could only differentiate the CFI group. Variance data are not displayed in the figure,but expressed in Table 2.

HCCFICFS

6

4

2

0

0 5 10 15Time (min)

D-A

rea

(cm

2 )

D-Area after secretin stimulation

pAll < 0.05 HCp CPI vs HC < 0.05

D-Area after secretin stimulation

CPICPS

6

4

2

0

D-A

rea

(cm

2 )

0 5 10 15

Time (min)

Figure 6. Figures displaying traced areas (median values) of the descending duodenum in the CF patients (left) and CP patients (right)compared to healthy controls. Measures were done before and in defined time intervals after secretin stimulation. Variance data are notdisplayed in the figure, but expressed in Table 2.

6 Engjom Trond et al.

Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y H

else

Ber

gen

- H

auke

land

uni

vers

itets

syke

hus

on 0

1/26

/15

For

pers

onal

use

onl

y.

the sufficient groups (p < 0.001). TheCPI group couldalso be differentiated from the CFI group (p < 0.001).There was a less prominent, but significant differencebetween the CFS group and the HC. There was a faircorrelation between both aspirated volume andD-bicarbonate against duodenal area in all groups.The correlation between aspirated volume and D-Area in the CF group is displayed in (Figure 7).

Diagnostic accuracy of duodenal area tracing

ROC curves were calculated using peak duodenalarea and AUC against exocrine pancreatic function.Sensitivity and specificity corresponded to suggestedcutoff values are displayed in Table III and an ROCcurve in the CF group in Figure 7. The diagnosticaccuracies using peak duodenal area and AUC werecomparable. The diagnostic accuracy was better in theCF groups compared to the CP groups.

Measures of gallbladder before and 15 min aftersecretin stimulation showed no emptying of the gall-bladder. There were no serious adverse effectsobserved. The administration of intravenous secretininduced only minor abdominal discomfort and sen-sation of hunger in some patients.

Relation to disease stage

Disease severity in the CP group was assessed by theMayo score (ranging from 0 to 15, with a score of 4 ormore diagnostic for CP). In this study, we defined ascore ranging from 4 to 6 as mild disease and score>6 as moderate to severe. This group comprises16 patients in the mild group and 23 patients in themoderate-to-severe group. The AUC calculated fromthe duodenal area measures was lower in the mod-erate-to-severe group compared to the mild CP group(58.0 cm2 [44.7–75.0] vs. 39.7 cm2 [22.3–52.9],p < 0.05). Variations within the groups were high

HCCFICFS

12

10

8

6

4

2

0

0 2 4 6 8 10 12 14 16 18 20

Aspirated volume (mL)

Pea

k D

-Are

a (c

m2 )

Peak D-Area vs Aspirated volume

r2 = 0.57p < 0.001

AUC = 0.97

ROC Curve CF vs HC

1.0

0.8

0.6

0.4

0.2

0.0

0.0 0.2 0.4 0.6 0.8 1.0

Sen

siti

vity

1-Specificity

Figure 7. Figure to the left display a scatter plot of the traced peak duodenal area against aspirated duodenal volume during endoscopic shorttest. Figure to the right display a receiver–operator curve of the same area data against exocrine functional status demonstrating a gooddiagnostic accuracy.

Table III. Diagnostic accuracy.

Group Sensitivity Specificity Cutoff (cm2) Accuracy

D-Area CF/HC 0.94 (0.70–1.0) 0.90 (0.77–0.97) 3.0 0.97D-Area CP/HC 0.69 (0.39–0.91) 0.73 (0.57–0.86) 4.7 0.73D-Area CF/CP/HC 0.76 (0.56–0.90) 0.76 (0.63–0.86) 3.9 0.83D-AreaAUC CF/HC 0.88 (0.62–0.98 0.93 (0.81–0.99) 25 0.93D-AreaAUC CP/HC 0.64 (0.31–0.89) 0.61 (0.44–0.76) 50 0.72

Table displays accuracy parameters for the duodenal area measurements Calculated from ROC curves, dependent of group selection. (Bestvalue with 95% CI).

Accuracy of secretin stimulated ultrasound 7

Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y H

else

Ber

gen

- H

auke

land

uni

vers

itets

syke

hus

on 0

1/26

/15

For

pers

onal

use

onl

y.

and diagnostic accuracy calculations for staging werenot performed.

Discussion

In this paper, we combined secretin-stimulated ultra-sonography with a 15-min EST. We address ultraso-nography measures of the pancreatic duct and DDafter secretin stimulation as a parameter for exocrinepancreatic volume output. Our study shows threemain findings: First, we found that ultrasonographyassessment of fluid filling in the descending part of theduodenum 5–15 min after secretin stimulation coulddifferentiate exocrine insufficient CF and CP patientsfrom exocrine sufficient patients and HC. Second, wewere able to differentiate curves representing ductaldilatation in our pancreatic insufficient CF patientsfrom the pancreas sufficient groups. However, wefound this parameter to be less clinically useful.Third, we found that by assessing descending duo-denal area instead of the diameter, we could improvediagnostic accuracy in all groups. Best results wereachieved in the assessment of exocrine pancreaticfunction in the CF group.In most cases, increased fluid after secretin admin-

istration gave considerable improvement of the visu-alization of the WD and duodenum.The correlation between exocrine failure and

reduced volume output after secretin stimulation inCFI patients has been established through tube-basedtests. It is estimated that CF patients secrete 40% lessfluid compared to non-CF controls [2,29]. Earlierstudies on secretin-stimulated ultrasonography havedemonstrated that changes in caliber of theWD coulddifferentiate patients with CP from healthy volunteers[21-23]. Several works are published on differentMRI-based methods on both ductal and duodenaldiameter and other methods of estimating duodenalfluid filling after secretin stimulation [3,5,30-32].These methods demonstrate correlation between exo-crine pancreatic function and secretin-stimulatedpancreatic fluid flow.To our knowledge, there are no publications on

using ultrasonography measures of fluid filling of theDD as a parameter of EPF. In this novel ultrasono-graphic method, we found that area measurements onthe fluid filled DD correlated well to exocrine func-tion testing and aspirated volumes in EST. Themethod was capable of identifying patients with themost severe pancreatic failure. Our results concerningWD diameter are comparable to earlier results withthe same method.Standardization of ultrasonography sections, image

quality and reproducibility are the main limitations ofthe method. A number of patients were excluded due

to poor ultrasonographic visualization. Standardiza-tion of the transducer position for measurements wasa challenge, partly explaining test variance within eachpatient series. The WD is a small structure; in CPpatients one of the disease features is caliber varia-tions. The time window of dilatation is short, andreversed by the opening of the sphincter of Oddi.These factors challenge the standardization of theWD measures, and probably explain why the param-eter is only suitable to detect the most severe volume-output failure and less useful in an individual diag-nostic setting. Choosing the exact anatomical locali-zation for measurement requires careful evaluation oflocal landmarks to reduce position errors. We did anindependent evaluation of the recorded data after thetest to improve standardization. From our experiencethe measurements on the WD is a challenging task toany ultrasonographers regardless of skills and experi-ence. The measurements on the DD are easier toperform. Changing to this parameter increases testfeasibility considerably. We argue that measures onstimulated fluid filling of the DD can be added indirect pancreas function testing in a practical way.Concerning the fluid flow in DD, we have not

compared our results to secretin MRI. Our ultra-sound method fails to detect some of the fluid flowingdistally in the duodenum. MRI can follow the fluidmore distally, contains a three-dimensional (3D)aspect and probably has fewer limitations in visuali-zation. Comparing the two methods needs furtherstudies. Measurements of small structures are vulner-able to operator bias. All examinations were carriedout by two doctors in agreement for the optimallocation for measurement. Operators were blindedto patient diagnosis and grade of pancreatic failure.Blinding to time interval and patient appearance wasnot possible. Repeated testing was not performed inthis protocol, thus data on interobserver variance arenot presented.One could argue that the differentiation between

very different diagnoses like CF and CP is not aclinical relevant question. However, we have includedpatient groups where due to differences in pathophys-iological mechanism we could expect different gradesof volume-output failure. Patient groups were alsodivided according to level of exocrine pancreaticinsufficiency. We chose cutoff limits for exocrinefailure for FE (FE1 < 100 mg/mg) and duodenalbicarbonate concentration (<80 meq/L). These cut-offs are according to a conservative diagnostic cutoffconsensus in CF [33], and were chosen to reduce therisk of false positives in the insufficient groups. Weargue that this creates uniform groups with differencesin expected volume-output failure spanning fromsevere failure in the CFI group, via the more

8 Engjom Trond et al.

Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y H

else

Ber

gen

- H

auke

land

uni

vers

itets

syke

hus

on 0

1/26

/15

For

pers

onal

use

onl

y.

acinar-dominated failure in the CP group toward onlymild or no failure in the CFS group and normalfunction in the HC group. We were able to identifyreduced volume outputs corresponding to expectedgrade of failure in pancreatic fluid flow in all pancreasinsufficient groups compared to HC. The demonstra-tion of the significantly better accuracy in the CFpopulation fits the hypothesis that ductal failure playsa larger role in CF- than in CP-exocrine failure.ERCP-based scores for the definition of severity inCP are not used any more. Good substitutes aremissing. In this study, we used the Mayo score asdiagnostic score. Although not validated for this pur-pose, we also used it as a severity defining score. Wefound that the precision of ultrasonography volume-output measures showed some correlation to thisseverity definition. High variance in the measuresreduces the feasibility of secretin ultrasonography asa good stage defining examination in CP.Lack of homogeneity and differences in severity of

the failure between the groups may still be confoun-ders explaining that the CP group presents a lowergrade of volume-output failure than the homogenousCF-group. There were some differences in age andsmoking habits between the CF and non-CF groupspossibly biasing the results, but we do not expect theeffect to alter our results significantly. A higher fre-quency of smoking and alcohol intake has beendescribed in several previous studies of CP patients[34,35].

Conclusion

Ultrasonography gives precise measurement of vol-ume transport in the DD and the WD and aftersecretin stimulation. We managed to identify patientswith severe pancreatic volume-output failure due toCF and CP compared to HC with good diagnosticaccuracy. The superiority of the method lies in itssafety and the possible clinical application of the directcombination with EST as an estimate of pancreaticfluid flow and exocrine pancreatic function. Giventhat MRI accessibility is limited and some subjects areunable to perform the MRI scanning, theultrasonography-based direct exocrine function testappears to be a simpler alternative in some patientgroups. Future research applying 3D ultrasono-graphic technology may improve the method.

Acknowledgements

The study was supported by MedViz , an interdisci-plinary research cluster from Haukeland UniversityHospital, University of Bergen and Christian

Michelsen Research AS. The Authors have receivedtravelling grants and scholarships from Abbott, theNorwegian Gastroenterology association and Norwe-gian cystic fibrosis foundation. A special acknowledge-ment to supporting lab technicians Liv Aasmul andAud Sissel Hjartholm for preserving and running anal-yses of duodenal juice and to Line Lærum for takingcare of the cystic fibrosis patients during the inclusion.

Declaration of interest: The authors report noconflicts of interest. The authors alone are responsiblefor the content and writing of the paper.

References

[1] Ramsey BW, Davies J, McElvaney NG, Tullis E, Bell SC,Drevinek P, et al. A CFTR potentiator in patients with cysticfibrosis and the G551D mutation. N Engl J Med 2011;365:1663–72.

[2] Kopelman H, Corey M, Gaskin K, Durie P, Weizman Z,Forstner G. Impaired chloride secretion, as well as bicarbon-ate secretion, underlies the fluid secretory defect in the cysticfibrosis pancreas. Gastroenterology 1988;95:349–55.

[3] Gillams AR, Lees WR. Quantitative secretin MRCP(MRCPQ): results in 215 patients with known or suspectedpancreatic pathology. Eur Radiol 2007;17:2984–90.

[4] Hellerhoff KJ, Helmberger H, Rosch T, Settles MR,Link TM, Rummeny EJ. DynamicMR pancreatography aftersecretin administration: image quality and diagnostic accu-racy. AJR Am J Roentgenol 2002;179:121–9.

[5] Sanyal R, Stevens T, Novak E, Veniero JC. Secretin-enhanced MRCP: review of technique and applicationwith proposal for quantification of exocrine function. AJRAm J Roentgenol 2012;198:124–32.

[6] Tiscornia OM, Cresta MA, de Lehmann ES, Celener D,Dreiling DA. Effects of sex and age on pancreatic secretion.Int J Pancreatol 1986;1:95–118.

[7] Wathle GK, Tjora E, Ersland L, Dimcevski G, Salvesen OO,Molven A, et al. Assessment of exocrine pancreatic functionby secretin-stimulated magnetic resonance cholangiopan-creaticography and diffusion-weighted imaging in healthycontrols. J Magn Reson Imaging 2013;39:448–54.

[8] Conwell DL, Zuccaro G, Morrow JB, Van LF,O’Laughlin C, Vargo JJ, et al. Analysis of duodenal drainagefluid after cholecystokinin (CCK) stimulation in healthyvolunteers. Pancreas 2002;25:350–4.

[9] Dreiling DA. Pancreatic secretory testing in 1974. Gut 1975;16:653–7.

[10] Kopelman H, Corey M, Gaskin K, Durie P, Weizman Z,Forstner G. Impaired chloride secretion, as well as bicar-bonate secretion, underlies the fluid secretory defect in thecystic fibrosis pancreas. Gastroenterology 1988;95:349–55.

[11] Lankisch PG. Function tests in the diagnosis of chronicpancreatitis. Critical evaluation. Int J Pancreatol 1993;14:9–20.

[12] Laugier R. Dynamic endoscopic manometry of the responseto secretin in patients with chronic pancreatitis. Endoscopy1994;26:222–7.

[13] Novis BH, Bornman PC, Girdwood AW, Marks IN. Endo-scopic manometry of the pancreatic duct and sphincter zonein patients with chronic pancreatitis. Dig Dis Sci 1985;30:225–8.

Accuracy of secretin stimulated ultrasound 9

Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y H

else

Ber

gen

- H

auke

land

uni

vers

itets

syke

hus

on 0

1/26

/15

For

pers

onal

use

onl

y.

[14] Lieb JG, Draganov PV. Pancreatic function testing: here tostay for the 21st century. World J Gastroenterol 2008;14:3149–58.

[15] Erchinger F, Engjom T, Tjora E, Hoem D, Hausken T,Gilja OH, et al. Quantification of pancreatic function using aclinically feasible short endoscopic secretin test. Pancreas2013;42:1101–6.

[16] Jensen NM, Larsen S. A rapid, endoscopic exocrine pancre-atic function test and the Lundh test: a comparative study.Pancreatology 2008;8:617–24.

[17] Stevens T, Conwell DL, Zuccaro G Jr, Lewis SA, Love TE.The efficiency of endoscopic pancreatic function testing isoptimized using duodenal aspirates at 30 and 45 min afterintravenous secretin. Am J Gastroenterol 2007;102:297–301.

[18] Stevens T, Conwell DL, Zuccaro G Jr, Van LF, Lopez R,Purich E, et al. A prospective crossover study comparingsecretin-stimulated endoscopic and Dreiling tube pancreaticfunction testing in patients evaluated for chronic pancreatitis.Gastrointest Endosc 2008;67:458–66.

[19] Tjora E, Wathle G, Engjom T, Erchinger F, Molven A,AksnesL,etal.Severepancreaticdysfunctionbutcompensatednutritional status in monogenic pancreatic disease caused bycarboxyl-ester lipase mutations. Pancreas 2013;42:1078–84.

[20] Erchinger F, Dimcevski G, Engjom T, Gilja O. Transab-dominal ultrasound of the pancreas: basic and new aspects.Imaging Med 2011;3:411–22.

[21] Bolondi L, Li BS, Gaiani S, Santi V, Gullo L, Barbara L.Impaired response of main pancreatic duct to secretin stimula-tion in early chronic pancreatitis. DigDis Sci 1989;34:834–40.

[22] Glaser J, Esser W, Holtmannspotter K. [Sonographic visu-alization of the pancreatic duct before and after secretinstimulation: an aid in the diagnosis of chronic pancreatitis].Ultraschall Med 1985;6:106–9.

[23] Glaser J, Hogemann B, Krummenerl T, Schneider M,Hultsch E, van HN, et al. Sonographic imaging of thepancreatic duct. New diagnostic possibilities using secretinstimulation. Dig Dis Sci 1987;32:1075–81.

[24] Layer P, Yamamoto H, Kalthoff L, Clain JE, Bakken LJ,DiMagno EP. The different courses of early- and late-onsetidiopathic and alcoholic chronic pancreatitis. Gastroenterol-ogy 1994;107:1481–7.

[25] Farrell PM, Rosenstein BJ, White TB, Accurso FJ,Castellani C, Cutting GR, et al. Guidelines for diagnosis of

cystic fibrosis in newborns through older adults: cystic fibrosisfoundation consensus report. J Pediatr 2008;153:S4–S14.

[26] Somogyi L, Ross SO, CintronM, Toskes PP. Comparison ofbiologic porcine secretin, synthetic porcine secretin, andsynthetic human secretin in pancreatic function testing.Pancreas 2003;27:230–4.

[27] WMA General Assembly. World medical association decla-ration of helsinki: ethical principles for medical researchinvolving human subjects 2013. Available from http://www.wma.net/en/30publications/10policies/b3. Last accessed June2014.

[28] Whittaker E, Robinson G. The trapezoidal and parabolicrules. the calculus of observations: a treatise on numericalmathematics. 4th edition. NewYork: Dover; 1967. p 156–88.

[29] Durie PR, Forstner GG. Pathophysiology of the exocrinepancreas in cystic fibrosis. J R Soc Med 1989;82:2–10.

[30] Bali MA, Sztantics A, Metens T, Arvanitakis M, Delhaye M,Deviere J, et al. Quantification of pancreatic exocrine func-tion with secretin-enhanced magnetic resonance cholangio-pancreatography: normal values and short-term effects ofpancreatic duct drainage procedures in chronic pancreatitis.Initial results. Eur Radiol 2005;15:2110–21.

[31] Erturk SM, Ichikawa T, Motosugi U, Sou H, Araki T.Diffusion-weighted MR imaging in the evaluation of pancre-atic exocrine function before and after secretin stimulation.Am J Gastroenterol 2006;101:133–6.

[32] Schneider AR, Hammerstingl R, Heller M, Povse N,Murzynski L, Vogl TJ, et al. Does secretin-stimulatedMRCP predict exocrine pancreatic insufficiency?:A comparison with noninvasive exocrine pancreatic functiontests. J Clin Gastroenterol 2006;40:851–5.

[33] Borowitz D, Baker SS, Duffy L, Baker RD, Fitzpatrick L,Gyamfi J, et al. Use of fecal elastase-1 to classify pancreaticstatus in patients with cystic fibrosis. J Pediatr 2004;145:322–6.

[34] Kristiansen L, Gronbaek M, Becker U, Tolstrup JS. Risk ofpancreatitis according to alcohol drinking habits: a -population-based cohort study. Am J Epidemiol 2008;168:932–7.

[35] Tolstrup JS, Kristiansen L, Becker U, GronbaekM. Smokingand risk of acute and chronic pancreatitis among women andmen: a population-based cohort study. Arch Intern Med2009;169:603–9.

10 Engjom Trond et al.

Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y H

else

Ber

gen

- H

auke

land

uni

vers

itets

syke

hus

on 0

1/26

/15

For

pers

onal

use

onl

y.