Design for 6 Sigma: The Utility of Process Capability Analysis Early in Pharmaceutical Process Development
Majuru, 1Lovy Pradeep, 1David Pereira, 2Christin Hollis, 2Harry Cocolas, 3Ray Thomas, 3Joseph Schlee
RESULTS CONTINUED
Solithromycin is currently in Phase 3 of clinical development for the treatment of community-acquired bacterial
pneumonia and uncomplicated gonorrhea. The dosage forms under development include hard gelatin capsules, a
lyophilized formulation for parenteral administration and a powder for oral suspension for the pediatric population.
The capsule dosage form consists of a solithromycin granulation with suitable excipients encapsulated into a size 0
hard gelatin capsule with a fill weight of 325 mg and a strength of 200 mg. In the course of development, clinical
supplies were successfully manufactured at batch sizes ranging from 5,000 to 150,000 capsules, and no major
process related issues were encountered. Based on the stage of clinical testing, the qualitative and quantitative
composition of the capsule formulation has already been fixed, however the manufacturing process needed to be
optimized.
In order to support the continued product development, it was necessary to manufacture batches for registration
stability testing. In preparation for registration batch manufacture, it was necessary to review the batch history and to
optimize the manufacturing process. In this regard, a Design-of-Experiments (DOE) approach was utilized to
characterize the design space and identify the operation ranges for parameters during the granulation, drying,
blending and encapsulation. A scale-up and three registration batches (160,000 capsules) were then manufactured
using the identified parameters.
The objective of this study was to utilize a Design-For-Six-Sigma (DFSS) approach and process capability analysis of
the scale-up and registration batches as a tool for an early assessment of the process’ potential for Six Sigma
Capability. This analysis will guide further development activities and validation batch manufacture.
100
95
Table 1: In-process and bulk product results for solithromycin capsules
Batch#1402287
Batch#1402288
Parameter
Acceptance Criteria
Batch
*1401679
In-process Testing Data
Xbar Chart
LSL
UCL=100.789
Overall
Within
Specifications
LSL
95
USL 105
LCL=98.509
3
4
5
6
7
8
9
10
90
93
96
R Chart
99
UCL=3.570
Blend Uniformity
Assay
Dissolution
NMT 2.5 % w/w
2.08
Mean
100.6
Std. Dev
0.81
Min
99.84
Max
101.98
Bulk Capsule Testing Data
90.0-110.0 % of Label Claim
100.8
Mean
97
NLT 75% Q in 45
RSD
1.3
minutes
Min
85
Max
98
NMT 6% w/w
3.0
USP/NF <905>
*Record for the blend that was encapsulated under record 1402067
3.1
1.90
1.62
1.70
99.41
0.42
98.85
100.23
99.05
0.43
97.95
99.53
99.56
0.27
99.22
99.94
100.1
96
1.0
95
97
3.3
97.9
95
1.5
93
96
3.3
99.1
94
1.1
93
95
3.3
2.4
3.4
2.1
108
3
4
5
6
7
8
9
98
100
102
Capability Plot
Overall
Within
StDev 0.7961
Cp
2.09
Cpk
1.95
PPM
0.00
100
415
20
10
Last 10 Subgroups
102
420
Within
5
Overall
StDev 0.7591
Pp
2.20
Ppk
2.04
Cpm
*
PPM
0.00
9
17
6
8
33
41
49
57
73
The in-process filled capsule weight checks for the scale-up and registration
batches indicated that the process was in control with all of the observations
within the control limits (Figure 4).
20
30
40
AD: 0.981, P: 0.014
UCL=19.88
_
R=8.71
10
9
17
25
33
41
49
57
65
73
410
Last 25 Subgroups
420
410
420
Overall
430
Within
Overall
StDev 4.270
Pp
3.28
Ppk
3.26
Cpm *
PPM 0.00
Specs
65
70
Batch #
f2
f1
65
6
1402287
76
3
Similar to clinical batch
based on f2 > 50
Not different from clinical
batch based on f1 < 15
70
Comments
Capability Plot
Within
StDev 4.206
Cp
3.33
Cpk
3.31
PPM 0.00
60
60
1402067
LCL=0
1
50
Normal Prob Plot
430
Figure 3: Process capability sixpack: blend uniformity
10
Figure 7: Dissolution profile in pH 2.5 phosphate buffer
Table 2: Comparison of Dissolution profiles with Clinical Batch
(14703.004) as the reference
385 396 407 418 429 440 451 462
0
10
Sample
0
Time (min)
Specifications
LSL 377.76
USL 461.73
LCL=413.61
65
Scale-up Batch 1402067
Reg. Batch 1402287
Reg. Batch 1402288
Reg. Batch 1402289
40
0
Overall
Within
R Chart
Specs
4
25
1
60
USL
__
X=419.95
1
LCL=0
2
2
Report Results
(% of Theoretical)
105
_
R=1.565
98
Moisture of dried
granulation
102
AD: 1.580, P: < 0.005
LSL
80
20
Capability Histogram
UCL=426.30
5
425
Normal Prob Plot
2
Batch # 1402288
Xbar Chart
USL
99
2
Batch # 1402289
Process Capability Sixpack Report for Filled Capsule Weight
__
X=99.649
4
Batch # 1402067
The process capability sixpack for the filled capsule weight is shown in Figure
6. As seen in the Figure, the process has a Cp of 3.33 and a Cpk of 3.31. This
indicates that the process is capable of producing capsules within
specifications and a fill weight distribution that is centered around the target.
S a m p le M e a n
100
1
Batch # 1402287
Capability Histogram
101
1
Batch
1402289
100
410
Figure 5: Boxplot for in-process filled capsule weight
The process capability sixpack is shown in Figure 3. Based on the results, the
Cp was 2.09 and the Cpk 1.95, indicating that the process is capable of
producing blends with a uniformity within 95-105 % of the theoretical value.
Dissolution profiles were also generated in pH 2.5 phosphate buffer using
USP apparatus I at 50 RPM for all the four batches and the profiles are
shown in Figure 7. The data indicates that the dissolution profiles of all the
four batches are comparable. The profiles of capsules from the B-M-E of
encapsulation were consistent with those of Figure 7. Furthermore, the
similarity of the profiles to a reference clinical batch was determined using f2
and f1 testing. Based on the analysis (Table 2), the dissolution profiles of all
the four batches are similar to the reference clinical batch (f2 >50). Also the
profiles are not different (f1<15).
420
Batch#1402289
Figure 2: Boxplots for blend uniformity for batches
1401679, 1402287, 1402288, and 1402289
0
Results
Batch
Batch
1402287
1402288
430
400
Batch#1401679
Process Capability Sixpack for Blend Uniformity
The manufacturing process yielded acceptable blends that were successfully encapsulated into size 0 hard gelatin
capsules. The in-process blend uniformity and capsules assay, moisture content and dissolution results are shown in
Table 1 below. The results indicated that the process was capable of producing blends that met the in-process
specifications and further produced capsules that met pre-established critical quality attributes. As shown in Table 1, the
results indicate that the granulation, drying, milling and blending processes yield in-process materials that meet
moisture and blend uniformity results pre-established criteria. Granulation moisture results ranged from 1.62 to 2.08 %
w/w while blend uniformity results for all four batches ranged from 97.95-101.98% of theoretical.
Moisture
Uniformity of Dosage units
(by weight variation)
Figure 1: Solithromycin
105
S a m p le R a n g e
Solithromycin, a new chemical entity under development by Cempra Pharmaceuticals, is a fourth-generation macrolide
antibiotic, and the first fluoroketolide. The chemical structure of solithromycin is shown in Figure 1 below.
440
90
Values
INTRODUCTION
RESULTS AND DISCUSSION
Filled Capsule Weight (mg) Boxplot
Boxplot of Blend Uniformity
V a lu e s
Conclusion: Process capability analysis indicated that the process used for registration batch manufacture was
capable of meeting specifications (Cp >1.33) and is centered with a Cpk greater than 1.33. The process has potential
for Six Sigma performance, thus demonstrating the utility of process capability analysis early in development.
Figure 5 shows a boxplot for in-process weight variation. Based on
the box plot, the median values are close to the theoretical value and
the spread of the data across the four batches is similar.
110
75
1402288
76
4
1402289
68
5
80
Sample
Figure 6: Fill weight capability sixpack
Run Chart for Batch # 1402067,1402287, 1402289, 1402288
450
Filled capsule weight (mg)
Results: The results for blend uniformity, assay, dosage uniformity and dissolution indicated that all batches
manufactured, were within specifications. Process capability analysis for the granulation and blending (blend
uniformity) as well as the encapsulation yielded Cp values greater than 1.33. Additionally, process analysis indicated
that it was centered as indicated by a Cpk greater than 1.33. The histogram and trend charts generated also indicated
that the process was capable of meeting specifications and that it was centered. Analysis of control charts indicated
that the encapsulation process was in control and that the process statistic was normal with at least 99.73% of all the
points within the control limits. Dissolution profiles of the batches indicated that they were consistent from batch-tobatch and with those of solithromycin capsule batches used in clinical testing.
The resulting blend was filled into Capsugel size 0 hard gelatin capsules using an Index K120i encapsulation machine.
After machine setup, capsules (10) were taken every 30 minutes and the individual gross filled capsule weight
determined. The target filled capsule weight was 420 mg with ±5% the upper and lower weight control limit across all
four batches. Capsules were weight-sorted using an IMA Precisa 120 weight sorter and inspected on an inspection belt.
Capsules were also submitted for bulk release testing, which included assay, content uniformity by weight variation,
moisture, and dissolution (single point). Additionally, dissolution profiles were collected for bulk capsules and for
capsules from the beginning, middle, and end (B-M-E) of encapsulation. Dissolution testing was conducted in 900 mL
(50 mM) of phosphate buffer with a pH of 2.5 at 37 ± 0.5 oC using USP apparatus 1 at 50 RPM. The data collected
was analyzed using Minitab® 17 Software. Process boxplots, probability plots, process control charts and the process
capability metrics (Cp and Cpk) were used to characterize the process. Additionally, the dissolution profiles of the
capsules manufactured were compared to those of capsules used in clinical studies.
Figure 2 shows the boxplot of the blend uniformity for the four batches. It is
noted that the scale-up batch (1401679) exhibited greater variability in blend
uniformity than the registration batches (1402287, 1402289 and 1402288).
% Solithromycin Released
METHODS
The manufacturing process involves high shear granulation, fluid bed drying, milling, blending and encapsulation.
Solithromycin and excipients, except for the magnesium stearate were screened through a 20 mesh screen and
transferred to a granulation bowl. Granulation was performed using a VG 200 high shear granulator with povidone
dissolved in water as the binder solution. Following granulation, the material was divided into four sub-lots of
approximately 25 kg each and dried using a fluid bed drier (GPCG-15) equipped with a 45 L bowl. The dried granules
were passed through an 18 mesh screen. The granules retained on the 18 mesh screen were milled using a Comil
equipped with a 0.5 mm round hole screen. All of the screened and milled dry granules were transferred to a V-blender
and the required amount of magnesium stearate was added followed by blending for three minutes. Ten samples for
blend uniformity testing were taken from the top middle and bottom of the blend.
Filled Capsule W eight (mg)
ABSTRACT
Purpose: Solithromycin is a fourth-generation macrolide antibiotic, and the first fluoroketolide. Current development is
focused on treatment of community-acquired bacterial pneumonia and uncomplicated gonorrhea. In preparation for
capsule registration batches it was necessary to optimize the manufacturing process. The objective of this study was
to utilize a Design-For-Six-Sigma (DFSS) approach and process capability analysis as a tool for an early assessment
of the process’ potential for Six Sigma Capability.
Methods: The manufacturing process involves high shear granulation, fluid bed drying, milling, lubrication and
encapsulation. Critical quality process parameters were identified, and a Design-of-Experiments (DOE) was used as a
guide towards understanding of the design space. Confirmation and registration batches were manufactured; blend
uniformity, in-process fill weight, assay and dissolution data were collected and analyzed using Minitab® 17 Software.
Process histograms, probability plots, process control charts and the process capability metrics (Cp and Cpk) were
used to characterize the process. Additionally, the dissolution profiles of the capsules manufactured were compared to
those of capsules used in clinical studies.
W4305
Inc., 6320 Quadrangle Drive, Suite 360, Chapel Hill, NC 27517
Pharma Solutions, 1100 Enterprise Drive, Winchester, KY 40391
Pharma Solutions, 160 Pharma Drive, Morrisville, NC 27560
Blen d U n ifirm ity ( % o f the o re tic al)
3Catalent
Mean Blend Uniformity (%)
1Cempra,
2Catalent
Sample Range
1Shingai
W4305
CONCLUSIONS
UCL = 440
mg
440
One scale-up and three registration scale batches (52 kg) were successfully manufactured and predetermined critical product attributes were met. Process capability
analysis indicated that the process used for registration batches manufacture was capable of meeting specifications (Cp >1.33) and is centered with a Cpk greater than
1.33. The capability of the process will continue to be monitored during scale-up to commercial batch size and the manufacture of validation and launch batches. The data
generated so far indicates that the process has potential for Six Sigma performance, thus demonstrating the utility of process capability analysis early in development.
430
420
410
LCL= 400 mg
400
390
1
5
10
15
20
25
30
35
40
45
50
55
60
Sample
Number of runs about median:
39
Expected number of runs:
41.0
Longest run about median:
8
Approx P-Value for Clustering: 0.326
Approx P-Value for Mixtures:
0.674
Number of runs up or down:
61
Expected number of runs:
53.0
Longest run up or down:
2
Approx P-Value for Trends:
0.984
Approx P-Value for Oscillation: 0.016
Figure 4: Filled capsule weight run chart
65
70
75
80
REFERENCES
1. Chowdhury, M.R., Process Capability Analysis in Pharmaceutical Production. International Journal of Pharmaceutical and Life Sciences Vol.2(2) 2013: 85-89