Sodium and potassium salts of bumetanide trihydrate: Impact of counterion on structure, aqueous solubility and dehydration kinetics

Article abstract of DOI:10.1039/c2ce06631a

A form of bumetanide potassium trihydrate that is structurally similar to bumetanide sodium trihydrate was identified. Structural analysis indicated however that the change from sodium to potassium salt resulted in meaningful changes in the packing arrangement and even greater changes on physicochemical properties. The potassium trihydrate salt was five times more soluble in water relative to the sodium trihydrate salt (27 mg mL^-1vs. 5.6 mg mL ^-1). Both salts underwent complete dehydration to their corresponding anhydrate forms following exposure to low humidity or elevated temperature. Arrhenius plots for the dehydration between 20-30 °C yielded a lower activation barrier for the potassium trihydrate compared to the sodium trihydrate (14.9 kcal mol^-1vs. 19.9 kcal mol^-1), translating into a 15-20-fold difference in the rate of dehydration. Such a 5.0 kcal mol^-1 gap suggests that the H₂O molecules are more tightly bound in the sodium trihydrate compared to potassium trihydrate. These differences in solubility and rate of dehydration further illustrate that changes to chemical composition can have a meaningful and unpredictable impact on the physicochemical properties of closely related pharmaceutical salt forms. The Royal Society of Chemistry 2012.

Full text of DOI:10.1039/c2ce06631a

View Article Online / Journal Homepage / Table of Contents for this issue

Dynamic Article Links <
C
CrystEngComm
Cite this: CrystEngComm, 2012, 14, 2428
www.rsc.org/crystengcomm
PAPER
Sodium and potassium salts of bumetanide trihydrate: Impact of counterion on
structure, aqueous solubility and dehydration kinetics†
*
Winston Ong,‡ Eugene Y. Cheung, Karen A. Schultz, Cartney Smith, James Bourassa and Magali B. Hickey
Received 6th December 2011, Accepted 10th February 2012
DOI: 10.1039/c2ce06631a
A form of bumetanide potassium trihydrate that is structurally similar to bumetanide sodium
trihydrate was identified. Structural analysis indicated however that the change from sodium to
potassium salt resulted in meaningful changes in the packing arrangement and even greater changes on
physicochemical properties. The potassium trihydrate salt was five times more soluble in water relative
to the sodium trihydrate salt (27 mg mLꢀ1 vs. 5.6 mg mLꢀ1). Both salts underwent complete dehydration
to their corresponding anhydrate forms following exposure to low humidity or elevated temperature.
ꢁ
Arrhenius plots for the dehydration between 20–30 C yielded a lower activation barrier for the
potassium trihydrate compared to the sodium trihydrate (14.9 kcal molꢀ1 vs. 19.9 kcal molꢀ1),
translating into a 15–20-fold difference in the rate of dehydration. Such a 5.0 kcal molꢀ1 gap suggests
that the H2O molecules are more tightly bound in the sodium trihydrate compared to potassium
trihydrate. These differences in solubility and rate of dehydration further illustrate that changes to
chemical composition can have a meaningful and unpredictable impact on the physicochemical
properties of closely related pharmaceutical salt forms.
Currently prescribed as the free acid in solid oral dosage and
Introduction
injectable forms, bumetanide has low aqueous solubility (0.1 mg
Utilization of salts is a common strategy to enhance the solubility
mLꢀ1),12 which subsequently limits its utility in low-volume
of low-molecular weight (MW) active pharmaceutical ingredi-
injectable delivery. Though bumetanide sodium trihydrate13 was
ents (APIs).1–4 In their neutral forms, ionizable APIs may suffer
independently determined to have an aqueous solubility of
from sub-optimal pharmaceutical properties, such as slow
5.6 mg mLꢀ1, which is significantly more soluble than the free
dissolution rate and low solubility in aqueous systems, often
resulting in poor oral bioavailability.5–9 In addition, drug
solubilities could facilitate the development of low-volume,
products that require reconstitution at the point of administra-
acid, identification of alternative salts with higher aqueous
parenteral formulations of bumetanide.
tion often depend on the identification of a high-solubility salt.
It has been shown recently that a change in counter-ion of
For this reason, APIs intended for parenteral administration
low-MW APIs can lead to a significant change in crystal
require extensive salt screening to ensure that a suitable form is
packing.14 In fact, the swapping of a sodium cation for potassium
selected for development.
Bumetanide10 (Fig. 1) is a potent loop diuretic with rapid onset
during salt formation was found to, more often than not, result in
significant differences in packing arrangements. While such anal-
and short duration of action. It is indicated to manage oedema
ysis illustrates that subtle variations in chemical composition can
typically associated with congestive heart failure, hepatic and
renal disease, and to treat mild to moderate hypertension.11
the change and the impact on the physicochemical properties of
be expected to change the packing arrangement, the magnitude of
the resultant materials remains highly unpredictable.15
Herein, the crystal structure and physicochemical properties of
Transform Pharmaceuticals, Inc. (A Johnson & Johnson company), 29
Hartwell Avenue, Lexington, MA 02421, USA
bumetanide potassium trihydrate (bumetanide K(H2O)3) are
reported. Though the structure was found to be similar in terms
† Electronic
Thermogravimetric and differential scanning calorimetric thermograms
of bumetanide and bumetanide Na(H2O)3, conversion between
supplementary
information
(ESI)
available:
K
trihydrate and anhydrate states of the sodium and potassium salts
under various humidities and temperatures, and raw solubility data of
bumetanide, bumetanide K(H2O)3 and bumetanide Na(H2O)3. CCDC
reference number 857276. For ESI and crystallographic data in CIF or
other electronic format see DOI: 10.1039/c2ce06631a
‡ Present address: Kala Pharmaceuticals, Inc., 135 Beaver St, Suite 309,
Waltham, MA, 02452 USA; Fax: +781-642-0399; Tel: +781-996-5255;
E-mail: winston.ong@kalarx.com
Fig. 1 Structure of bumetanide and bumetanide salts.
This journal is ª The Royal Society of Chemistry 2012
2428 | CrystEngComm, 2012, 14, 2428–2434

of space group and overall packing arrangement to the corre-
sponding sodium salt trihydrate (bumetanide Na(H2O)3),13 the
structures were neither identical nor isostructural. The physico-
chemical properties of the sodium and potassium trihydrate salts
drastically differ as follows: (i) the potassium salt is five times
more soluble in water at pH > 7.5 (ii) the potassium salt
dehydrates 15–20 times faster, and (iii) the sodium salt is the only
trihydrate that is physically stable under ambient conditions
(22 ꢁC and 30–50% RH). Crystal structure analysis was con-
ducted to provide insight into the aforementioned difference in
solubility, dehydration rate and physical stability.
were acquired under ambient conditions (22 ꢁC and 30–50% RH)
at a power setting of 46 kV at 40 mA in reflection mode, while
oscillating about the omega-axis from 0–5ꢁ at 1ꢁ/s and spinning
about the phi-axis at 2ꢁ/s. Exposure time was 5 min. The
diffractograms were integrated over 2-theta from 2–40ꢁ and chi
(1 segment) from 0–360ꢁ at a step size of 0.02ꢁ using the cylint
utility in the RINT Rapid display software provided with the
instrument. The dark counts value was set to 8 as per the system
calibration; normalization was set to average; the omega offset
was set to 180ꢁ; and no chi or phi offsets were used for the
integration. Variable temperature and variable humidity PXRD
data were obtained on a Bruker D8 Advance X-ray diffractom-
eter with Sol-X detector equipped with a Parr TTK-450 and
MRI-700900 chamber, respectively. Samples were scanned from
3.5–40ꢁ (0.02ꢁ/step and 0.65 s/step) at 45 kV and 40 mA with
Methods and materials
Bumetanide was obtained from AK Scientific (Union City, CA).
Sodium hydroxide and potassium hydroxide were purchased
from Aldrich (St. Louis, MO) as 1.0 M solutions. Solvents were
HPLC grade and were purchased from EMD Chemicals
(Gibbstown, NJ).
ꢀ
Cu-Ka radiation (1.54 A).
Single crystal X-ray diffraction
The specimen chosen for X-ray diffraction study was a clear,
colorless needle measuring 0.049 ꢂ 0.077 ꢂ 0.573 mm. Single
crystal data were collected at 100 K on a Bruker kappa APEX II
diffractometer system supplied with a Mo-Ka fine-focus sealed
Preparation of the bumetanide salts
Potassium salt anhydrate (bumetanide K). Aqueous potassium
hydroxide (13.3 mL ꢂ 1.0 M) was added drop wise to a stirred
solution containing bumetanide (4.89 g, 13.3 mmol) dissolved in
methanol (150 mL) at ambient temperature. After stirring the
solution for 5 min, the solvent was evaporated under reduced
pressure to produce the anhydrous salt as a white solid. Isolated
yield: 5.37 g (11.8 mmol, 89%).
ꢀ
X-ray tube (l ¼ 0.71069 A) and operated at 50 kV, 30 mA.
Frames were collected with a scan width of 0.5ꢁ in u and an
exposure time of 20 s/frame. A total of 2180 frames were inte-
grated. The total data collection was 24 h. The crystal structure
was solved by direct methods and refined by least squares using
CRYSTALS.16 Hydrogen atoms were located by Fourier
difference or calculated in idealized positions. The refinement
converged at R ¼ 0.0360, Rw ¼ 0.0697, for 7515 independent
reflections with I $ 2s and 262 variables.
Sodium salt trihydrate (bumetanide Na(H2O)3). This salt was
prepared in the same manner as bumetanide K, except that
sodium hydroxide (13.3 mL ꢂ 1.0 M) was used in place of
potassium hydroxide. Isolated yield starting from bumetanide
free acid (4.91 g, 13.3 mmol) was 5.80 g (13.2 mmol, 99%). The
crystal structure of this salt has previously been deposited in the
Cambridge Structural Database as WINJAE.13
Dynamic Vapor Sorption (DVS)
Vapor sorption–desorption data of approximately 10 mg of
bumetanide salt were acquired using DVS-1 (Surface Measure-
ment Systems, Allentown, PA). The instrument was allowed to
equilibrate for at least an hour at each temperature prior to the
start of the experiment. Humidity steps were ramped in 10%
increments after the change in mass (dM/dT) fell under 0.02%
within a ten-minute window. The activation energy (Ea) for the
removal of H2O from the crystal lattices of bumetanide Na(H2O)3
and bumetanide K(H2O)3 was determined using the classical
Arrhenius method,17 which consisted of isothermal dehydration
experiments performed over a series of temperatures (approxi-
Potassium salt trihydrate (bumetanide K(H2O)3) single crystals.
Bumetanide K was dissolved in H2O at 50 ꢁC to generate a 50-mg
mLꢀ1 solution, which was then allowed to spontaneously
equilibrate to ambient temperature. Needle-shaped crystals of
bumetanide K(H2O)3 started growing after 24 h.
Sodium salt anhydrate (bumetanide Na). Exposure of 15–25 mg
of bumetanide Na(H2O)3 to 0% relative humidity (RH, at 25 ꢁC)
or >80 ꢁC in the variable humidity or variable temperature
chambers of a Bruker D8 Advance diffractometer (see below)
generated the anhydrate salt in situ, as observed by the emergence
of new PXRD peaks.
ꢁ
ꢁ
ꢁ
ꢁ
mately 30 C, 27.5 C, 25 C and 20 C). The initial humidity
setting for the Na salt corresponded to an absolute vapor pres-
sure equivalent to 30% P/P0 at 25 ꢁC, where P is the actual vapor
pressure and P0 is the saturated vapor pressure. The initial
humidity setting for the K salt was 90% P/P0 at all temperatures.
Dehydration was initiated by switching the initial humidity
setting to 100% N2 (i.e., 0% P/P0). Zero-order rates of dehy-
dration (k) were extracted from data points covering 0–75%
dehydration using the least squares method. Each dehydration
measurement was obtained in triplicate. Least squares analysis of
the corresponding Arrhenius plots (ln(k) vs. Tꢀ1) generated
slopes equal to ꢀEa/R, where R is the universal gas constant.
Powder X-ray diffraction (PXRD)
High-resolution PXRD patterns were obtained using the D/Max
Rapid X-ray Diffractometer equipped with a copper source
ꢀ
(Cu/Ka 1.5418 A), manual x–y stage, and 0.3-mm collimator.
Samples were loaded into a 0.3-mm boron rich glass capillary
tube by sectioning off one end of the tube and tapping the open,
sectioned end into a bed of sample. The loaded capillary was
mounted in a holder secured into the x-y stage. Diffractograms
This journal is ª The Royal Society of Chemistry 2012
CrystEngComm, 2012, 14, 2428–2434 | 2429

Solubility
The pH-solubility profiles starting from bumetanide K and
bumetanide Na(H2O)3 were determined from buffer-free
suspensions stirred at ambient temperature (22 ꢁC) for 24 h.
Drug substance was suspended in de-ionized water (1 mL) in
a 3-mL vial equipped with a magnetic stir bar and pH was
adjusted by titrating the suspensions with 0.1 M HCl or the
appropriate hydroxide base (0.1 M KOH or 0.1 M NaOH). After
stirring the suspension for 24 h, the pH was measured, undis-
solved solids were separated from the mixture via filtration using
a 0.2-mm PTFE membrane, and the solution was diluted
2–1000-fold for LC (liquid chromatography) analysis. The
starting concentrations of the suspensions varied according to
the target pH as follows: at pH < 5, the initial concentrations
were 6 mg mLꢀ1 for both Na and K salts, while at pH > 5, the
initial concentrations were 20 mg mLꢀ1 and 50 mg mLꢀ1 for the
Na and K salts, respectively.
Fig. 3 DVS gravimetry of bumetanide potassium and sodium salts at
25 ꢁC. Filled and unfilled symbols represent sorption and desorption
cycles, respectively. Each anhydrate form showed an uptake of three H2O
molecules upon hydration. The theoretical weight increase for formation
of trihydrates from potassium and sodium anhydrates are 13.4% and
14.0%, respectively. Changes in mass are relative to the respective
anhydrous forms.
Liquid chromatography
Bumetanide content was determined using a Waters ACQUITY
Ultra Performance LC system equipped with a PDA detector. A
2.1 mm ꢂ 50 mm BEH C18 column with particle size of 1.7 mm
was used and maintained at 40 ꢁC. Mobile phase A consisted of
0.1% formic acid in H2O, while mobile phase B was 0.1% formic
acid in CH3CN. The flow program consisted of a two-minute
monitored at 90% RH by PXRD (Fig. S2†), new reflections of
bumetanide K(H2O)3 emerged that were consistent with the
PXRD data recorded when bumetanide K was suspended in H2O
(Fig. S3†), verifying that identical forms were produced when
bumetanide K was exposed to conditions where water activity is
high (e.g., suspension in bulk water and exposure to 90% RH). A
hydrated bumetanide potassium salt has been previously
disclosed,19 but in the absence of sufficient characterization data,
no further information is available to determine the stoichiom-
etry of waters of hydration. It is unclear whether an intermediate
bumetanide potassium monohydrate was formed upon desorp-
tion of the trihydrate at 50% RH (monohydrate is theoretically
4.5% heavier than bumetanide K).
gradient of 60% to 10% mobile phase A at 0.7 mL minꢀ1
.
Analysis was performed with a 5-mL injection, and quantitation
was based on an extracted wavelength of 268 nm.
Results and discussion
Characterization and physical stability of bumetanide salts
TGA confirmed that the isolated sodium salt was a trihydrate
(bumetanide Na(H2O)3, Fig. S1†), which is in agreement with the
reported crystal structure.13 DVS measurements showed that
bumetanide Na(H2O)3 converted to anhydrous bumetanide Na
upon evacuation from the humid atmosphere (i.e., #10% RH),
but quantitatively recovered the 14% mass loss18 upon rehydra-
tion to RH $ 20% (Fig. 3). Both bumetanide sodium and
bumetanide potassium formed trihydrate salts, but each
trihydrate salt differed in the humidity bracket at which the
hydrates were physically stable (RH > 20% for the sodium salt
and RH > 40% for the potassium salt).
The PXRD patterns of bumetanide and its various salt forms are
shown in Fig. 2. The isolated potassium salt, bumetanide K, was
found to be anhydrous using thermogravimetric analysis (TGA,
Fig. S1†). DVS gravimetric measurements displayed a 13% mass
uptake of bumetanide K from 30% RH to 90% RH, signifying the
formation of bumetanide potassium trihydrate, bumetanide
K(H2O)3.18 This trihydrate salt completely dehydrated upon
reversing the RH to #40% (Fig. 3). When form conversion was
TGA also showed a 12.2% decrease in mass of bumetanide
Na(H2O)3 at >80 ꢁC (Fig. S1†), indicating the formation of
anhydrous bumetanide Na at elevated temperꢁatures.18 The
PXRD of bumetanide Na(H2O)3 acquired at 120 C or 0% RH
(Fig. S2†) displayed similar reflections, further demonstrating
that the trihydrate converts to the same anhydrous form at low
humidity or elevated temperature.
Dehydration kinetics of trihydrate salts
The activation energy (Ea) for the dehydration of bumetanide
K(H2O)3 and bumetanide Na(H2O)3 was extracted from the
Arrhenius plots using DVS.17 Dehydration was initiated by
switching the humid environment surrounding the trihydrate
Fig. 2 PXRD patterns of bumetanide and bumetanide salts. Patterns of
the trihydrate salts are shown in red.
2430 | CrystEngComm, 2012, 14, 2428–2434
This journal is ª The Royal Society of Chemistry 2012

salts to a dry N2 atmosphere. Fig. 4 shows the Arrhenius plots
ꢁ
values correspond to the trihydrate forms. Despite the increased
solubility of the potassium salt, disproportionation of the salt to
the free acid was not observed for suspensions stored at pH > 7.5
and 22 ꢁC for 8 weeks. Given that the risk of salt dispropor-
tionation increases with higher solubility salts, the observations
reported here bear early evidence that the potassium salt
possesses the promising solution stability required to formulate
and Table 1 lists the isothermal k values between 20–30 C and
the corresponding Ea parameter. Dehydration of both salts
followed zero-order kinetics at all temperatures and the Ea value
of the potassium salt was lower than the sodium salt by 5.0 kcal
molꢀ1 20
The dehydration rate of bumetanide K(H2O)3 was
.
notably 15–20-fold faster than bumetanide Na(H2O)3, which is
likely due to the 5.0-kcal molꢀ1 difference in Ea. The elevated Ea
barrier for dehydration of the sodium salt also suggests that the
H2O molecules are bound more tightly to the crystal lattice of
bumetanide Na(H2O)3. Such rationale is consistent with the DVS
profiles in Fig. 3, which show that a lower RH is required to
dehydrate bumetanide Na(H2O)3.
a
physiologically-compatible,23 slightly alkaline parenteral
product. At pH < 4, the sodium and potassium salts dis-
proportionated into the free acid, causing the insoluble bume-
tanide to precipitate. The equilibrated forms at pH 5–7 were not
determined, though it is reasonable to assume that the equili-
brated suspensions consisted mixtures of the salt and free acid.
pH-dependent solubility
Single-crystal structures of bumetanide K(H2O)3 and
bumetanide Na(H2O)3
The pH-solubility21 profile of bumetanide in aqueous solutions at
22 ꢁC is summarized in Table S1† and Fig. 5. As predicted,
solubility is highest at pH regions where the carboxylic acid of
bumetanide (pKa ¼ 3.7)22 is >99% ionized. Surprisingly, the
solubility of the potassium salt at pH > 7.5 is five times greater
compared to the sodium salt (27–30 mg mLꢀ1 vs. 5–6 mg mLꢀ1).
At the conclusion of the solubility measurements, PXRD anal-
ysis of the equilibrated solids in the suspension verified that the
undissolved drug substance at pH > 7.5 was either bumetanide
Na(H2O)3 or bumetanide K(H2O)3, confirming that the solubility
The crystallographic data for the sodium and potassium trihy-
drate salts are listed in Table 2. A close examination of the
PXRD patterns in Fig. 2 and the crystal data would indicate
structural similarity because the unit cells show similar cell
lengths and volumes, and both crystallize in monoclinic space
group C2/c. Though the PXRD patterns additionally result in
high similarity scores when compared using PXRD similarity
tools, visual inspection of the patterns reveals clear differences in
the two salt forms. Therefore, while it is striking that two forms
Fig. 4 Left: Time dependence for the zero-order dehydration of bumetanide K(H2O)3 upon exposure to dry N2 in the DVS. Lines show the best fit to the
zero-order kinetics. Dehydration beyond 50% is not shown due to crowding of data points. Right: Arrhenius plots for the dehydration of bumetanide
Na(H2O)3 and bumetanide K(H2O)3. Each point represents the average of triplicate measurements between 20 and 30 ꢁC. Lines show best fit to the
Arrhenius equation shown in the inset. The corresponding Ea values and zero-order rates are listed in Table 1.
Table 1 Zero-order rates (k) and activation energies (Ea) for the dehydration of bumetanide Na(H2O)3 and bumetanide K(H2O)3. k values were obtained
in triplicate. Ea was extracted from the Arrhenius plots in Fig. 4
Dehydration reaction
T/ꢁC
k (% minꢀ1
)
Ea (kcal molꢀ1
14.9 ꢃ 1.1
)
bumetanide K(H2O)3 / bumetanide K
19.7
25.1
27.6
30.1
19.9
25.0
27.7
30.0
5.04 ꢃ 0.05
7.99 ꢃ 0.19
9.90 ꢃ 0.24
11.7 ꢃ 0.10
0.252 ꢃ 0.01
0.527 ꢃ 0.02
0.653 ꢃ 0.04
0.775 ꢃ 0.07
bumetanide Na(H2O)3 / bumetanide Na
19.9 ꢃ 3.3
This journal is ª The Royal Society of Chemistry 2012
CrystEngComm, 2012, 14, 2428–2434 | 2431

Fig. 5 Plot of 24-h concentration vs. pH shows a marked increase in
solubility of the potassium salt relative to the sodium salt at pH > 7.5.
Inset shows the equilibrated forms at pH < 4 and pH > 7.5 determined by
PXRD. Symbols represent the starting forms (squares: bumetanide K,
circles: bumetanide Na(H2O)3) used at the start of the experiment. Raw
data are listed in Table S1.†
13
6 Overlay of bumetanide Na(H2O)3 (blue) and bumetanide
Fig.
K(H2O)3 (gray) shows a nearly exact superimposition of the bumetanide
in the two salts.
carboxylic acid of bumetanide, and one O atom belonging to the
sulfonamide moiety of bumetanide, respectively. In bumetanide
Na(H2O)3, each six-coordinate Na center is linked to neigh-
boring Na cores by two pairs of bridging H2O molecules. The
remaining two coordination sites are occupied by a non-bridging
H2O molecule and one O atom of the carboxylic moiety of
bumetanide. The greater number of bridging H2O molecules in
bumetanide Na(H2O)3 provides a network of H-bonds which is
considerably different to that found in bumetanide K(H2O)3. This
network of cross-linked H2O molecules is perhaps responsible
for the aforementioned discrepancy in DEa ¼ 5.0 kcal molꢀ1 for
dehydration, and even possibly the five-fold solubility enhance-
ment of the potassium trihydrate. Moreover, an analysis of the
packing coefficients with removal of water and cation from the
two crystal structures indicates that even though bumetanide
K(H2O)3 starts with a higher packing coefficient than bumetanide
Na(H2O)3, the percentage of void space after the loss of all three
H2O molecules and potassium cation differs only by 1% to that
of bumetanide Na (Table 3). Purely from a structural perspective,
bumetanide K(H2O)3 will show more voids when it sheds water,
and this looser packing during dissolution may contribute in
some way to its increased solubility.
having identical hydration states and sharing similar single-
crystal structures but drastically differing in physical properties,
the crystallographic data indicate that the reduced unit cells are
not isostructural, and variations in crystal packing must exist.
The most prominent difference in the two structures is in the
arrangement of the H2O and bumetanide molecules around the
potassium and sodium cations. These subtle differences in
structure provide insight into the molecular basis for the
dramatic differences in the material properties.
As the conformations of bumetanide in the two crystal struc-
ꢀ
tures overlap nearly exactly (root mean squared misfit ¼ 0.02 A,
Fig. 6), the primary structural differences in the crystals are solely
due to the placement of the cations and water molecules, likely
a consequence of the larger size of the potassium cation
compared with the sodium cation. The subtle differences in
arrangement are evident in the packing diagrams (Fig. 7).
In bumetanide K(H2O)3, the K center is six-coordinated, but
only one of the four H2O molecules forms a bridge with
a neighboring K core. The remaining two positions are occupied
by one O atom from the carbonyl group belonging to the
Table 2 Crystallographic data for bumetanide Na(H2O)313 and bumetanide K(H2O)3
Compound
Bumetanide Na(H2O)3
Bumetanide K(H2O)3
Chemical formula
Formula weight (g molꢀ1
T/K
C17 H19 Na1 N2 O5 S1$3H2O
440.44
293
1.5418
C17 H19 K1 N2 O5 S1$3H2O
456.56
100
0.71073
)
ꢀ
Wavelength (A)
Crystal size/mm
Crystal system
Space group
0.5 ꢂ 0.4 ꢂ 0.2
0.049 ꢂ 0.077 ꢂ 0.573
Monoclinic
C2/c
Monoclinic
C2/c
ꢀ
Unit cell dimensions (A)
a ¼ 44.819(2)
a ¼ 45.797(2)
b ¼ 5.3639(8)
b ¼ 5.4342(3)
c ¼ 17.694(4)
c ¼ 17.5936(9)
b ¼ 95.91(2)ꢁ
b ¼ 102.160(3)ꢁ
3
ꢀ
Volume/A
4231.2(9)
4280.3(4)
Z
8
1.38
8
1.42
Density (calculated) g cmꢀ3
Final R indices
I > 3.0s(I) R1 ¼ 0.058, wR2 ¼ 0.059
I > 2.0s(I) R1 ¼ 0.036, wR2 ¼ 0.070
2432 | CrystEngComm, 2012, 14, 2428–2434
This journal is ª The Royal Society of Chemistry 2012

Fig. 7 Differences in crystal packing (left ¼ bumetanide Na(H2O)3,13 right ¼ bumetanide K(H2O)3.
Table 3 Analysis of the packing coefficients with removal of water and cation from bumetanide Na(H2O)313 and bumetanide K(H2O)3
Bumetanide Na(H2O)3 (WINJAE)13 (%)
Bumetanide K(H2O)3 (%)
Trihydrate + metal ion
ꢀ1 water molecule
69.3
67.5
66.1
64.0
56.3
73.6
72.6
71.0
68.7
55.6
ꢀ2 water molecules
ꢀ3 water molecules
ꢀall water molecules and metal ion
Conclusions
References
1 M. L. Peterson, M. B. Hickey, M. J. Zaworotko and O. Almarsson, J.
Pharm. Pharm. Sci., 2006, 9, 317–326.
2 A. T. M. Serajuddin, Adv. Drug Delivery Rev., 2007, 59, 603–
616.
3 S. L. Childs, G. P. Stahly and A. Park, Mol. Pharmaceutics, 2007, 4,
323–338.
4 P. H. Stahl and C. G. Wermuth, ed., Handbook of pharmaceutical
salts: properties, selection, and use, Wiley-VCH, New York,
2002.
5 C. Han and B. Wang, ed., Factors That Impact the Developability of
Drug Candidates: An Overview, in Drug Delivery: Principles and
Applications, John Wiley & Sons, Inc., Hoboken, 2005.
6 C. A. Lipinski, F. Lombardo, B. W. Dominy and P. J. Feeney, Adv.
Drug Delivery Rev., 1997, 23, 3–25.
7 L.-F. Huang and W.-Q. Tong, Adv. Drug Delivery Rev., 2004, 56, 321–
334.
8 C. R. Gardner, C. T. Walsh and O. Almarsson, Nat. Rev. Drug
Discovery, 2004, 3, 926–934.
A new potassium salt trihydrate (bumetanide K(H2O)3) that is
five times more soluble than the known bumetanide Na(H2O)3
form (27–30 mg mLꢀ1 vs. 5–6 mg mLꢀ1) is reported. The
dehydration of the sodium salt is 15–20-fold slower than the
potassium salt due to a more energetic activation barrier (DEa ¼
5.0 kcal molꢀ1). Perhaps, the differentiated H-bond distances and
crystallographic position of H2O molecules illustrate that an
obvious switch in countercations can, and more often than not,
lead to subtle changes in structure but meaningful and unpre-
dictable variations in physicochemical properties. These data
also highlight the importance of utilizing crystallography to
further understand the effects of changing counterions towards
the physical properties of API salts. Finally, the solubility
increase for the potassium salt, which did not show evidence of
disproportionation to the free acid at pH regions (>7) where
solubility exceeded 25 mg mLꢀ1, can be particularly important to
low-volume parenteral formulations of bumetanide, as it facili-
tates the preparation of concentrated drug solutions at the near-
neutral pH.
9 M. N. Martinez and G. L. Amidon, J. Clin. Pharmacol., 2002, 42,
620–643.
10 C. T. Supuran, Curr. Pharm. Des., 2008, 14, 641–648.
11 A. Ward and R. C. Heel, Drugs, 1984, 28, 426–464.
12 H. W. Nielsen, E. Bechgaard, B. Twile, E. Didriksen and
G. T. Almtorp, Pharm. Dev. Technol., 2001, 6, 145–149.
This journal is ª The Royal Society of Chemistry 2012
CrystEngComm, 2012, 14, 2428–2434 | 2433

13 M. Saviano, R. Fattorusso, A. Lombardi, L. Zaccaro and C. Pedone,
Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 1995, C51, 395–398.
14 P. A. Wood, M. A. Oliveira, A. Zinkb and M. B. Hickey,
CrystEngComm, 2012, DOI: 10.1039/C2CE06588F.
15 J. F. Remenar, J. M. MacPhee, B. K. Larson, V. A. Tyagi, J. H. Ho,
D. A. McIlroy, M. B. Hickey, P. B. Shaw and O. Almarsson, Org.
Process Res. Dev., 2003, 7, 990–996.
19 P. W. Feit, (Loevens Kemiske Fabrik Produktionsaktieselskab, Den.).
U S 13,806,
Application: US, 1976, pp. 19 pp Division of
p. 534
20 Crystal defects and particle size can influence activation energy. Both
factors were not taken into consideration in this study.
21 The suspensions used in solubility measurements were titrated
with aqueous NaOH or KOH to adjust pH. Suppression of
solubility due to common-ion effect, if any, was not accounted
for.
16 P. W. Betteridge, J. R. Carruthers, R. I. Cooper, K. Prout and
D. J. Watkin, J. Appl. Crystallogr., 2003, 36, 1487.
17 J. Y. Khoo, J. Y. Y. Heng and D. R. Williams, Ind. Eng. Chem. Res.,
2010, 49, 422–427.
22 B. Song, A. K. Galande, K. Kodukula, W. H. Moos and S. M. Miller,
Drug Dev. Res., 2011, 72, 315–426.
18 Theoretical mass changes: bumetanide K to bumetanide K(H2O)3 ¼
+13.4%, bumetanide Na to bumetanide Na(H2O)3 ¼ +14.0%, and
bumetanide Na(H2O)3 to bumetanide Na ¼ ꢀ12.3%.
23 Parenteral products are generally formulated between pH 4 and 9.
For examples, see R. G. Strickley, PDA J. Pharm. Sci. Technol.,
1999, 53, 324–349.
2434 | CrystEngComm, 2012, 14, 2428–2434
This journal is ª The Royal Society of Chemistry 2012