Formulation study for orally disintegrating tablet using partly pregelatinized starch binder.

Abstract

In this study, we aimed to design orally disintegrating tablets by employing a formulation design approach that enables the production of such tablets in the same facilities used for the production of solid dosage forms on an industrial scale. First, we examined the relationships between the types of binders used in the tablets and the properties of orally disintegrating tablets prepared by the wet granulation method. Results revealed that partly pregelatinized starch is a relatively suitable binder for orally disintegrating tablets as it also serves as a disintegrant. Next, we employed a central composite design for 2 factors, namely, corn starch and partly pregelatinized starch, in order to design granules suited for orally disintegrating tablets composed of D-mannitol, corn starch or partly pregelatinized starch. The effects of these 2 factors on 3 types of responses, namely, 50% granule size, compressing index and disintegrating index, were analyzed with a software package, and responses to changes in the factors were predicted. This study investigated the effects of binder type and binder content in orally disintegrating tablets, and provided evidence that the binder exerts a strong influence on tablet properties, and is therefore an important component of orally disintegrating tablets.

Full text

Page 1

August 2011959Regular Article
Orally disintegrating tablets are widely accepted in current
medical practice as patient-friendly alternatives to conven-
tional oral solid dosage forms, such as regular tablets and
capsules, especially by those who have problems or difficul-
ties swallowing (e.g. elderly and pediatric patients, and pa-
tients with cerebrovascular event) and those requiring fluid
restriction. Zydis technology, developed by Cardinal Health
in the early 1980s, is said to be the origin of the orally disin-
tegrating tablet technology. There are currently 3 classes of
orally disintegrating tablet technology, namely, crystalliza-
tion control,1—4)process control5—7)and formulation de-
sign.8,9)Characteristically, technologies for both types of
crystallization control and process control improve the disin-
tegrating speed of the tablets. However, the resulting tablets
tend to be too soft, and special additives included in the for-
mulations, or special production machinery and complex
procedures are required for their production. Therefore, high
manufacturing costs and low productivity are problems asso-
ciated with these 2 classes of technologies. On the other
hand, tablets that are based on the formulation design show
reasonable hardness, albeit with slightly lower disintegrabil-
ity compared with those of the other 2 technologies. Further,
the formulation design technology enables production of
orally disintegrating tablets with conventional tableting ma-
chinery, and thus contributes to cost effectiveness and stable
production. Recently, formulations containing crospovi-
done10)or corn starch11)as a disintegrant, a saccharide or a
disaccaride as a filler,12)sodium stearyl fumarate13)or a su-
crose ester of a fatty acid14)as a lubricant were reported. In
addition, formulation optimization using design of experi-
ments (DOE)15—17)and improvements in manufacturing
processes18)has been demonstrated. Accordingly, the quality
of orally disintegrating tablets is continuously improving.
In this study, we focused on formulation design technol-
ogy, and evaluated various types of binders used in the wet
granulation process. D-Mannitol (Man) and corn starch (CS)
were respectively used as a filler and a disintegrant of orally
disintegrating tablets. We also employed DOE to analyze and
identify formulations containing a suitable binder. Because
the study aimed to construct a model applicable to the for-
mulation of orally disintegrating tablets with consideration of
the ethical aspects involved in testing disintegration of tablets
in humans, only placebo formulations that lacked inclusion
of drugs were examined.
Experimental
Materials
Shoji Foodtech Co., Ltd. (Tokyo, Japan). Corn starch (CS) was purchased
from Nippon Starch Chemical Co., Ltd. (Osaka, Japan). Partly pregela-
tinized starch (PGS: Starch 1500) was purchased from Colorcon Japan, Llc.
(Shizuoka, Japan). D-Sorbitol (Sor: Sorbogen) was obtained from SPI
Pharma, Inc. (Wilmington, U.S.A.). Macrogol 4000 (Mac) was obtained
from NOF Corp. (Tokyo, Japan). Hypromellose (Hyp: TC-5E) was pur-
chased from Shin-Etsu Chemical Co., Ltd. (Tokyo, Japan). Hydroxypropyl-
cellulose (HPC: HPC-SL) was purchased from Nippon Soda Co., Ltd.
(Tokyo, Japan). Povidone (PVP: Povidone K30) was obtained from ISP Ltd.
(Wayne, U.S.A.). Light Anhydrous Silicic Acid (LASA: Adosolidar 101)
was purchased from Freund Co., Ltd. (Tokyo, Japan). Calcium Stearate
(CaSt) was purchased from Nitto Kasei Kogyo Co., Ltd. (Kanagawa, Japan).
All other chemicals were of regent grade.
Preparation of Tablets
1)Preparation of Tablets with Different Types
of Binder: Various binders were used for tablet formulations (Table 1), and
the tablets were prepared as follows. Each binder was dispersed or dissolved
in purified water to make a binding solution. Man was placed in a fluid-bed
granulator (LAB-1, Powrex Corp., Hyogo, Japan), and granules were pre-
pared while spraying a binding solution mist. This was followed by sizing
with a powermill (P-02S, Dalton, Tokyo, Japan) with a screen size of
1.5mm in diameter. CS, CaSt and LASA were then added to the resulting
granules in the ratio shown in Table 1, and mixed by hand for 2min in a
plastic bag. The mixture was set in a rotary tablet press (Correct 12HUK,
Kikusui Seisakusho Ltd., Kyoto, Japan) and compressed into tablets using
the following settings: punch and die diameter of 8mm, curvature of 12R,
compression force of 4, 8 or 12kN and tablet weight of 200mg.
2) Preparation of Tablets Containing Granules Prepared According to a
Central Composite Design: Granule formulations containing CS, PGS and
Man were prepared using a central composite design for 2 variables, CS and
PGS (Table 2). A dispersion of PGS in purified water was used as a binder
solution. Man and CS were placed in the fluid-bed granulator, and the gran-
D-Mannitol (Man: Mannit P) was purchased from Mitsubishi
Formulation Study for Orally Disintegrating Tablet Using Partly
Pregelatinized Starch Binder
Kazuki MIMURA,a,bKen KANADA,bShinya UCHIDA,aMasaki YAMADA,band Noriyuki NAMIKI*,a
aDepartment of Pharmacy Practice and Science, School of Pharmaceutical Sciences University of Shizuoka; 52–1 Yada,
Suruga-ku, Shizuoka 422–8526, Japan: and bPharmaceutical Research & Development, Kissei Pharmaceutical Co., Ltd.;
19–48 Yoshino, Matsumoto, Nagano 399–8710, Japan.
Received January 28, 2011; accepted May 6, 2011; published online May 17, 2011
In this study, we aimed to design orally disintegrating tablets by employing a formulation design approach
that enables the production of such tablets in the same facilities used for the production of solid dosage forms on
an industrial scale. First, we examined the relationships between the types of binders used in the tablets and the
properties of orally disintegrating tablets prepared by the wet granulation method. Results revealed that partly
pregelatinized starch is a relatively suitable binder for orally disintegrating tablets as it also serves as a disinte-
grant. Next, we employed a central composite design for 2 factors, namely, corn starch and partly pregelatinized
starch, in order to design granules suited for orally disintegrating tablets composed of D-mannitol, corn starch or
partly pregelatinized starch. The effects of these 2 factors on 3 types of responses, namely, 50% granule size,
compressing index and disintegrating index, were analyzed with a software package, and responses to changes in
the factors were predicted. This study investigated the effects of binder type and binder content in orally disinte-
grating tablets, and provided evidence that the binder exerts a strong influence on tablet properties, and is there-
fore an important component of orally disintegrating tablets.
Key words
orally disintegrating tablet; binder; partly pregelatinized starch; hardness; disintegration time; central composite design
Chem. Pharm. Bull. 59(8) 959—964 (2011)
© 2011 Pharmaceutical Society of Japan
∗ To whom correspondence should be addressed.e-mail: namiki@u-shizuoka-ken.ac.jp

Page 2

ules were prepared while spraying a binding solution mist. Granules were
then filtered through a No. 18 mesh sieve, and 490g of the resulting size-
controlled granules were mixed with 2.5g of LASA and 7.5g of CaSt with a
V-type blender (DV-1, Dalton, Tokyo, Japan) for 2min. The mixture was set
in the rotary tablet press to make tablets using the following settings: punch
and die diameter of 8mm, curvature of 12R, compression force of 4, 6, 8, 10
or 12kN and tablet weight of 200mg.
Viscosity of Binder Solution
The viscosity of each binding solution
(6% w/w) and slurry was measured with a rotating viscometer (VISCONIC
EMD, Toki Sangyo Co., Ltd., Tokyo, Japan) at 100rpm at 20°C.
Granule Characteristics
Granules were mesh controlled and their
weight-standard distributions were obtained with a sonic sifter (L-3PS,
Seishin Enterprise Co., Ltd., Tokyo, Japan). The 50% granule sizes were
then calculated using a regression equation. Bulk density was measured with
a powder tester (PT-N, Hosokawa Micron Corp., Osaka, Japan).
Tablet Characteristics
Hardness of tablets was measured with a hard-
ness tester (TS-75N, Okada Seiko Co., Ltd., Tokyo, Japan). Five tablets were
tested for each formulation, and the mean hardness was calculated. Oral dis-
integration time was examined both in humans and with an orally disinte-
grating tablet tester (ODT-101, Toyama Sangyo Co., Ltd.). Five healthy men
were asked to take a tablet and roll it gently with their tongue in the oral cav-
ity. The time required for tablet disintegration was measured, and the mean
time for the 5 individuals was obtained. The orally disintegrating tablet
tester was preliminarily evaluated to find the conditions that correlate with in
vivo results in humans. Five tablets of each formulation were tested with the
tester under pre-determined conditions (beaker temperature: 37?2°C; rotat-
ing speed: 140rpm; weight: 20g; screen size: 2.0mm in diameter), and the
mean value was obtained.
Statistical Analysis
tablet hardness, and tablet hardness and disintegration time, were observed,
the slope of the linear regression curve for each correlation was defined as
follows. 1) Compressing index (N/kN): the slope of the regression line of
tablet hardness on compression force. A higher index value indicates that an
incremental increase in tablet hardness, in response to an increase in com-
pression force, will be harder. In other words, formulations with a high
index value indicate good compressibility. 2) Disintegrating index (s/N): the
slope of the regression line of oral disintegration time on tablet hardness. A
lower index value indicates that tablet disintegrability is sufficient regardless
of tablet hardness. JMP6 (SAS Institute Japan Ltd., Tokyo, Japan) software
was used to analyze the effects of 2 factors (CS and PGS) on 3 responses
(50% granule size, compressing index and disintegrating index) by perform-
ing analysis of variance (ANOVA) tests, and to predict factors and responses
involving second-order interactions.
Since correlations between compression force and
Results and Discussion
Three tablet-making methods are used in formulation de-
sign technology. The first is the direct compression method
that produces tablets by directly compressing a mixture of
raw materials. The second is the dry granulation method,
which involves preparation of dry granules from raw materi-
als before tablet compression. The third is the wet granula-
tion method, where granules are made from raw materials
with the aid of a binder solution prior to tablet compression.
Modification of materials is easier in the third method than in
the other two; therefore, flexible control of compressibility
and disintegrability, by altering the manufacturing condi-
tions, is its distinctive feature. A binder is added as a solution
or slurry in the wet granulation method and helps with the
formation of interparticle bonds and growth of granules. It
also enhances the hardness of the resulting tablets.
Many studies on the formulation design of orally disinte-
grating tablets have been reported, and many of them investi-
gated fillers, disintegrants and lubricants. With respect to
binders, compounds that do not strongly enhance interparti-
cle bond strength, such as solutions of a carbohydrate or a
sugar alcohol19)and a slurry of CS or crospovidone,20)were
previously tested. Low interparticle bond strength correlates
with good disintegrability of tablets, but also associates with
tableting problems and deterioration in tablet hardness,
which are not desirable properties for industrial-scale pro-
duction. Amongst the widely used binders are cellulose and
its derivatives, starch and its derivatives and synthetic poly-
960Vol. 59, No. 8
Table 1.Formulation of Tablets Using Different Types of Binders
Formulation (mg)
FunctionComponent
A-1A-2A-3A-4 A-5A-6 A-7
Filler Man172.48172.48 172.48172.48172.48 172.48 172.48
Binder Man
PGS
Sor
Mac
Hyp
HPC
PVP
3.52
—
—
—
—
—
—
—
3.52
—
—
—
—
—
—
—
3.52
—
—
—
—
—
—
—
3.52
—
—
—
—
—
—
—
3.52
—
—
—
—
—
—
—
3.52
—
—
—
—
—
—
—
3.52
Disintegrant
Lubricant
Glidant
Total
CS
CaSt
LASA
2020 202020 2020
3
1
3
1
3
1
3
1
3
1
3
1
3
1
200200200200200200200
Table 2.
Central Composite Design
Formulation of Granules Using Partly Pregelatinized Starch with
Variable level
in formulation
Actual value
Formulation
CSPGSCS (%) PGS (%)Man (%)
B-1
B-2
B-3
B-4
B-5
B-6
B-7
B-8
B-9
B-10
B-11
?1
?1
?1
51
1
5
5
3
3
6
0
3
3
3
94
80
90
76
75
95
82
88
85
85
85
1 19
?11
1
0
0
2
2
0
0
0
5
1
2
2
0
0
0
0
0
19
22
ø—
?ø—
2
ø—
12
12
12
12
12
?ø—
Center points of formulations were repeated three times to estimates the experimen-
tal error (B-9, B-10 and B-11).

Page 3

mers, all of which demonstrate diverse functions and charac-
teristics. In the present study, we compared a wide range of
binders of orally disintegrating tablets to identify one that en-
hances interparticle bond strength to an adequate level for in-
dustrial-scale production, and also produces good disintegra-
bility of tablets. Binders enhance interparticle bond strength,
while decreasing disintegrability of tablets in a dose-depen-
dent fashion. Therefore, it is important to examine carefully
the types and concentrations of binders for designing formu-
lations of orally disintegrating tablets.
Binder for Orally Disintegrating Tablet
Table 1, 7 formulations were prepared using 7 types of
binders, and tablets were prepared from each formulation by
the wet granule method. A compression force of 4, 8 or
12kN was used. The viscosity of each binder and the 50%
granule size of granules are shown in Table 3. The viscosity
of the binder solution containing HPC was the highest, while
that of the binder solutions containing Man, Sor and Mac
were the lowest. Size distribution of the tablet material is an
important factor influencing the tableting process. When the
granule size of the material is too small, tableting problems,
attributed to poor flowability and compressibility, tend to
occur. On the other hand, excessively large-sized granules
cause deterioration in filling properties and compression fail-
ure. Therefore, tablet material suitable for industrial-scale
production needs to have an appropriate granule size.
The relationship between compression force and tablet
hardness is shown in Fig. 1a. Regardless of the types of for-
mulation, tablet hardness increases in line with increased
compression force. Hardness of tablets containing PVP as a
binder was the most responsive to increases in compression
force, while that of tablets containing Man was the least re-
sponsive. It was also shown that tablet hardness is easily in-
creased by increasing the compression force, when PGS,
Hyp, HPC and PVP are used as the binder. It is considered
that the highest hardness of tablets containing PVP resulted
from the binding strength of PVP itself and lower viscosity
of binder solution enabled uniformly-distributed granules.21)
The relationship between tablet hardness and disintegration
time is shown in Fig. 1b. For a given tablet hardness, tablets
containing Hyp are the most difficult and slowest to disinte-
grate, while those containing Man are the easiest and fastest
to disintegrated. Because the hardness of tablets containing
Man was insufficient, we performed a separate experiment in
an attempt to improve that hardness. However, we encoun-
tered tableting problems, and could not produce Man-con-
taining tablets with sufficient hardness. We estimate that the
criteria for suitable characteristics of orally disintegrating
tablet are both ?50N for hardness and ?40s for disintegra-
As shown in
tion time. Among the binders tested, Sor, HPC and PVP as
well as PGS showed acceptable characteristics. In addition,
PGS appeared to be the better profile than others since it con-
sistently and independently adds good disintegrability to
tablets to increase tablet hardness. It is generally believed
that the disintegrating time of immediate-release tablets and
orally disintegrating tablets increases as their hardness in-
creases. Our results have demonstrated that an increase in
tablet hardness has a different extent of impacts on disinte-
grating time depending on the type of binder used. Taken to-
gether, it was shown that PGS can be a suitable binder for the
production of orally disintegrating tablets because it pro-
duces granules with appropriate size and contributes to good
compressibility and tablet disintegrability. A scanning elec-
tron microscope image of a granule containing PGS is shown
in Fig. 2. As assumed, CS and Man were observed as rod-like
and spherical particles, respectively, and PGS was observed
as irregular particles connecting Man and CS. In addition to
its role as a binder, PGS is known to enhance tablet disinte-
grability by its ability to induce swelling.22)It has been sug-
gested that PGS-containing tablets make suitable orally dis-
integrated tablets, because both PGS and CS contribute to
good disintegrability.
Central Composite Design for Granules of Orally Dis-
integrating Tablet
Granules containing CS, PGS and Man
were prepared according to the formulations shown in Table
2. A second-order model of the rotatable central composite
design with 2 factors, namely CS and PGS, wherein the cen-
ter points run 3 times, was used. Five levels of compression
force, 4, 6, 8, 10 and 12kN, were tested for tableting. As
August 2011 961
Table 3. Viscosity of Binder Solution and 50% Granule Size
Formulation
Viscosity
(mPa·s)
50% Granule size
(mm)
A-1
A-2
A-3
A-4
A-5
A-6
A-7
2.4
7.0
2.6
2.8
16.8
36.5
5.5
72
116
90
87
190
173
132
Fig. 1.
tween Hardness and Disintegration Time with ODT-101 (b)
Results are expressed as mean?S.D. (n?5). A-1 (?, Man), A-2 (?, PGS), A-3 (?,
Sor), A-4 (?, Mac), A-5 (?, Hyp), A-6 (?, HPC), A-7 (?, PVP).
Relation between Compression Force and Hardness (a), and be-
Fig. 2.
tinized Starch as a Binder
Scanning Electron Micrograph of a Granule Using Partly Pregela-

Page 4

shown in Table 2, although Man, as well as PGS and CS, can
be considered as variables, Man was used as a filler and its
content did not vary largely depending on the type of formu-
lation. Therefore, we did not include this component in the
factors of the central composite on the assumption that it has
only minor influences on compressibility and disintegrability.
The relationship between compression force and tablet
hardness is shown in Fig. 3a, and the relationship between
tablet hardness and disintegration time is shown in Fig. 3b.
The dependence of the increase in tablet hardness on the in-
creases in compression force varied across the formulations.
Increases in compression force made marked improvements
in the hardness of tablets made from some formulations,
while small improvements in the hardness of tablets were
made from the others. Nevertheless, hardness of all formula-
tions increased linearly with increasing compression force,
albeit at different rates. With respect to disintegration time,
increases in tablet hardness correlated with longer tablet dis-
integration time, regardless of formulation type. The extent
of the influence of an increase in hardness on disintegration
time varied across the formulations. Some formulations
showed little increases in disintegration time in line with in-
creases in hardness, while others showed marked increases.
As shown in Fig. 1b, tablets containing binders such as Hyp
showed a sudden large increase in disintegration time when
tablet hardness reached a certain level. When PGS was used
as a binder, the disintegration time of the tablets increased
linearly with increases in hardness, independent of the PGS
concentration in the formulation. On the basis of the above
findings, we defined the slope of the regression line of tablet
hardness on compression force and that of tablet disintegra-
tion time on tablet hardness as the compressing index and the
disintegrating index, respectively. We used these findings as
indexes for designing orally disintegrating tablets. A high
compressing index and a low disintegrating index are charac-
teristics of formulations suitable for orally disintegrating
tablets. The 50% granule size, compressing index and disin-
tegrating index of each tested formulation are shown in Table
4. The 50% granule size varied widely from 46 to 217mm.
The compressing index ranged from 2.28 to 8.58, while the
disintegrating index ranged from ?0.01 to 1.13.
Formulation Analysis Using Design of Experiments
The effects of the factors (CS and PGS) on the 50% granule
size, compressing index, and disintegrating index, were ana-
lyzed using JMP6 statistical analysis software, and the results
of ANOVA and sorted effect estimates are shown in Table 5,
Table 6 and Table 7, respectively. The p values for the above
responses were 0.0017, 0.0038 and 0.0019, respectively, indi-
cating that 3 responses change depending on the amounts of
CS and PGS. The sorted effect estimates showed that the in-
tercept and PGS had the strongest impacts, and second-order
CS had effects on the 50% granule size. The intercept and
PGS influenced the compressing index, while CS, PGS and
second-order CS impacted on the disintegrating index. The p
values for these factors were less than 0.05, indicating that
the effects of these factors on corresponding responses were
962Vol. 59, No. 8
Fig. 3.
tween Hardness and Disintegration Time in Oral Cavity (b)
Results are expressed as mean?S.D. (n?5). B-1 (?), B-2 (?), B-3 (?), B-4 (?), B-
5 (?), B-6 (?), B-7 (?), B-8 (∗), B-9 (—), B-10 (—), B-11 (—).
Relation between Compression Force and Hardness (a), and be-
Table 4.
of Formulations Using Partly Pregelatinized Starch with Central Composite
Design
50% Granule Size, Compressing Index and Disintegrating Index
50% Compressing
index
(N/kN)
Disintegrating
index
(s/N)
Formulation Granule size
(mm)
B-1
B-2
B-3
B-4
B-5
B-6
B-7
B-8
B-9
B-10
B-11
89
76
5.23
4.62
7.55
7.51
5.99
6.95
8.58
2.28
6.60
6.19
6.60
0.19
0.01
1.01
0.32
0.16
1.13
0.53
?0.01
0.17
0.13
0.20
203
175
140
143
217
46
123
122
106
Table 5. ANOVA and Sorted Effect Estimates for 50% Granule Size
Analysis of variance
Source
Model
Error
C. total
DFSum of squares
27141.948
382.961
27524.909
Mean square
5428.39
76.59
F ratio
70.8738
Prob?F
0.0017*
5
5
10
Sorted effect estimates
Term
Intercept
CS (%)
PGS (%)
CS (%)?PGS (%)
CS (%)?CS (%)
PGS (%)?PGS (%)
Estimate
117.01397
?5.603535
55.235294
?3.75
12.074886
6.5013623
S.E.
5.045095
3.078527
3.001809
4.375851
3.657529
3.363939
t ratio
23.19
?1.82
18.4
?0.86
3.30
1.93
Prob?|t|
?0.001*
0.1284
?0.001*
0.4306
0.0214*
0.1111
∗p?0.05.

Page 5

significant. Contour plots depicting the relationships of CS
and PGS with 3 responses are shown in Fig. 4. While Tables
5, 6 and 7 present the mathematically obtained predicted
response values, the contour plots provide visual and simple
information on factor-dependent changes in responses. Con-
tour lines run vertically in the 50% granule size and com-
pressing index plots, indicating that these responses tend to
be influenced by PGS. On the other hand, contour lines run
diagonally in the disintegrating index plot, indicating that
this response is under the influence of PGS and CS.
In this study, we established the effectiveness of PGS as a
binder of orally disintegrating tablets. Furthermore, we em-
ployed DOE and formulation analysis to examine properties
of tablets in connection with their formulation, and success-
fully obtained effective information for designing orally dis-
integrating tablets. On the basis of the findings of this study,
we intend to perform further studies to optimize the formula-
tion containing active ingredients, and furthermore, the tablet
production procedure.
August 2011963
Fig. 4. Contour Plots for 50% Granule Size (a), Compressing Index (b) and Disintegrating Index (c)
Table 6. ANOVA and Sorted Effect Estimates for Compressing Index
Analysis of variance
Source
Model
Error
C. total
DF Sum of squares
27.309701
1.622790
28.932491
Mean square
5.46194
0.32456
F ratio
16.8289
Prob?F
0.0038*
5
5
10
Sorted effect estimates
Term
Intercept
CS (%)
PGS (%)
CS (%)?PGS (%)
CS (%)?CS (%)
PGS (%)?PGS (%)
Estimate
6.4735639
?0.250126
1.7247059
0.1425
0.0544584
?0.417566
S.E.
0.328415
0.200399
0.195405
0.28485
0.23809
0.218979
t ratio
19.71
?1.25
8.83
0.50
0.23
?1.91
Prob?|t|
?0.001*
0.2672
0.0003*
0.6381
0.8281
0.1148
∗p?0.05.
Table 7.ANOVA and Sorted Effect Estimates for Disintegrating Index
Analysis of variance
Source
Model
Error
C. total
DFSum of squares
1.4250435
0.0624475
1.4874909
Mean square
0.285009
0.012489
F ratio
22.8199
Prob?F
0.0019*
5
5
10
Sorted effect estimates
Term
Intercept
CS (%)
PGS (%)
CS (%)?PGS (%)
CS (%)?CS (%)
PGS (%)?PGS (%)
Estimate
0.1638771
?0.279116
0.2282353
?0.1275
0.2204247
0.030113
S.E.
0.064424
0.039312
0.038332
0.055878
0.046705
0.042956
t ratio
2.54
?7.10
5.95
?2.28
4.72
0.70
Prob?|t|
0.0517
0.0009*
0.0019*
0.0714
0.0052*
0.5146
∗p?0.05.