Clinica Chimica Acru, 203 (1991) 109-118
109
0 1991 Elsevier Science Publishers B.V. All rights reserved 0009-8981{91/$03.50
CCA 05124
Effect of simvastatin on the apparent size of LDL particles in patients with type IIB hyperlipoproteinemia Shui P. Zhao ‘, Leny Hollaar Augustinus
‘, Ferdinand M. van ‘t Hooft H.M. Smelt *, Jan A. Gevers Leuven 3
‘,
and Arnbud van der Laarse ’ Departments of ’ Cardiology and ’ Internal Medicine, University Hospital, Leiden, and ’ Gaubius Institute TNO, Leiden (The Netherlands)
(Received 17 July 1990; revision received 19 August 1991; accepted 21 August 1991) Key words: Low density lipoprotein; LDL particle size; Type IIb hyperlipoproteinemia;
Simvastatin
Summary After 15 weeks of simvastatin therapy (20 mg/day), low density lipoprotein particle size in sera of 16 patjents with type I,‘b hyperlipoproteinemia increased significantly from 233 f 5.0 A to 237 f 7.0 A (P < 0.051, analyzed by 2-16% polyacrylamide gradient gel electrophoresis. Under simvastatin therapy the concentrations of total cholesterol, total triglyceride, very low density lipoprotein cholesterol and triglyceride, low density lipoprotein cholesterol and apolipoprotein B in serum fell significantly by 30%, 30%, 43%, 28%, 36% and 26%, respectively, and the concentration of high density lipoprotein cholesterol rose significantly by 14%. The changes of low density lipoprotein particle size induced by simvastatin therapy were correlated best with the changes of very low density lipoprotein triglyceride concentration (r2 = 0.438, P < 0.01). Our results suggest that simvastatin therapy, additionally to a reduction of the serum cholesterol concentration, increases low density lipoprotein particle size which may contribute to reduction of the risk of coronary heart disease in patients with type IIb hyperlipoproteinemia. Introduction Low density lipoprotein (LDL) isolated from human plasma contains approximately 50% cholesterol (free and esterified), 25% protein, 20% phospholipid and
Correspondence to: Dr. A. van dei Laarse, Department of Cardiology, Building 1. CS-P24, University Hospital, Rijnsburgenveg 10, 2333 AA Leiden, The Netherlands.
110
5% triglyceride. The heterogeneity of LDL with respect to size and density has been demonstrated by several groups [l-5]. Using 2-16% polyacrylamide gradient gel electrophoresis, Krauss and Burke [2] have demonstrated the presence of seven distinct LDL bands. They showed that these bands of LDL are associated with different LDL densities by means of density gradient ultracentrifugation. Recently, McNamara et al. [l] found that LDL subfraction distribution is influenced by gender and plasma lipoprotein levels. The characteristics of LDL subfractions as described by them were comparable to those observed after gradient ultracentrifugation [3]. Furthermore, Crouse et al. [4] reported a higher frequency of small LDL in patients with premature coronary artery disease than in healthy individuals. A recent study on LDL subclass patterns supports this observation [5]. It is now agreed that the heterogeneity in physical (size and density), chemical, and immunological properties of LDL is an inherent characteristic of circulating LDL [1,5-71. Although it has been proposed that the distribution of LDL subfractions is determined genetically [7-91, several studies have demonstrated that the concentrations of lipids in serum, especially triglyceride, influence the migration distance and patterns of LDL bands on polyacrylamide gradient gel electrophoresis [1,2,5] and LDL subfraction distribution after gradient ultracentrifugation [3]. The purpose of the present study is to examine whether LDL particle size is affected by simvastatin therapy in patients with type IIb hyperlipoproteinemia and if so, whether therapy induced changes of LDL particle size are related to changes in serum lipid levels. Materials and methods Patients
Sixteen patients (11 male and 5 female) with type IIb hyperlipoproteinemia were studied from November 1988 to January 1990. The ages of the patients ranged from 37 to 66 years, with an average age of 53.4 yr. The diagnosis of type IIb hyperlipoproteinemia was based on a serum total cholesterol concentration > 8 mmol/l, serum total triglyceride concentration > 2 mmol/l, LDL-cholesterol concentration > 4 mmol/l, and very low density lipoprotein (VLDL) cholesterol concentration > 1 mmol/l. Thirteen patients had either a positive family history of hyperlipoproteinemia or premature development of atherosclerosis, and were considered to have familial combined hyperlipoproteinemia. Three patients had xanthoma with a positive family history of hyperlipoproteinemia or premature development of atherosclerosis, and were considered to have familial hypercholesterolemia. None of the selected patients was overweight, nor was any of them drinking alcohol (> l/2 ounce/day). None had renal, thyroid or liver disease, or diabetes. All patients were given a diet containing fat (30% of total calories) with a sat : mono : poly ratio of 1: 1: 1, and < 300 mg cholesterol per day for at least 6 wk. Then, the patients started to receive simvastatin (20 mg/day) therapy.
111
Methods In all patients venous blood samples were obtained before taking simvastatin and after 15 wk (average) of simvastatin therapy. Blood samples were taken in the morning after > 12 h of fasting. After clotting, serum was obtained by centrifugation at 1,344 x g for 10 min at 20°C < 4 h after sampling. Five ml of fresh serum was ultracentrifuged with 5 ml of normal saline in a Kontron 75.13 rotor at 412,000 X g for 15 h at 15°C (Kontron AG, Zurich, Switzerland). Supernatant contained VLDL, and infranatant contained LDL and high density lipoprotein (HDL). HDL-cholesterol was determined by precipitating LDL in the infranatant by using phosphotungstic acid and MgCI,. LDL-cholesterol was the cholesterol measured in the infranatant minus HDL-cholesterol. In serum and ultracentrifugal fractions the triglyceride and cholesterol concentrations were measured enzymatitally using test kits (Boehringer, Mannheim, FRG). LDL-apolipoprotein (APO) B in the ultracentrifugal infranatant was measured by immunonephelometry (Beckman, Brea, California, USA). The remaining serum was stored at -20°C. The two serum samples from each patient obtained before and during therapy were analyzed simultaneously in the same polyacrylamide gel. Each sample was analyzed in two gels, one stained with Sudan black B which interacts with fatty moieties of lipoprotein, and a second stained with Coomassie Brilliant blue which interacts with protein moieties of lipoprotein. The method of gradient gel electrophoresis was according to McNamara et al. [l]. Briefly, each lane of the 2-16% nondenaturating polyacrylamide gels (Pharmacia LKB, Uppsala, Sweden) ‘was loaded with 7.5 ~1 of mixed serum prepared by mixing three volumes of serum with one volume of a solution of 1.17 mol/l sucrose and 0.1% bromophenol blue. Pre-electrophoresis of samples at 70 V for 20 min at 8°C was followed by electrophoresis at 450 V for 6 h at 8°C. One gel was stained with Sudan black B for 15 h. The Sudan black B staining solution was prepared by adding 0.8 g of Sudan black B and 4.0 g of zinc acetate to 300 ml of 60% ethanol in distilled water and heating to boiling for 10 min with slow stirring. Then the staining solution was filtered hot using filter paper. After cooling to room temperature the solution was refiltered before use. The gel was destained with a 50% solution of Cellosolve (Sigma, St. Louis, MO, USA) in distilled water for 2 h. The second gel was stained with Coomassie Brilliant .blue for 2 h. This staining solution was prepared by adding 1.25 g of Coomassie Brilliant blue R-250 to 250 ml of methanol, and 300 ml of 2.91 mol/l acetic acid in distilled water. The gel was destained for 15 h with the same solution except for Coomassie Brilliant blue. After destaining, the gels were rehydrated in a solution of 25% Cellosolve in distilled water to restore gel size and shape. Stained gels were scanned on a LKB 2202 Ultrascan Laser Densitometer (LKB, Paramus, NJ, USA). In each gel, the reference standard serum obtained from a healthy volunteer was applied to lane 1, 6, and 11 of a total of 11 lanes. When gel length was 70 mm after rehydration (original length of gels ‘is 78-79 mm), the migration distance of reference standard LDL was 14 mm. Meanwhile, a set of standard proteins with known hydrated diameters (HMW electrophoresis calibration kit, Pharmacia,
112
Piscataway, NJ, USA) such as thyroglobulin (170 A>, ferritin (122 A>, and catalase (104 A> was run in a number of gels and stained with Coomassie Brilliant blue to calibrate the diameter of LDL particles.
Statistical analysis Results are expressed as means and standard deviations (mean f SD). Differences between mean values were statistically evaluated by paired Student’s z test. Correlation analysis and stepwise multiple regression analysis were performed using SPSS/PC +@ (SPSS Inc., Chicago, IL, USA) software.
1
I
Fig. 1. Gradient black B (bottom).
1
10
2
3
4
5
6
7
6
9
2
3
4
5
6
7
6
9
polyacrylamide Both gels were
slab gels stained with Coomassie loaded
identically
with normal
10
Brilliant
11
I
11
I
blue (top) or with Sudan
serum (lanes
1, 6 and II)
patient sera (lanes 2, 3, 4, 5, 7, 8, 9, IO). Sera were applied to top of each gel.
and with
113
Results Electrophoresis of serum on 2-16% polyacrylamide gradient gels produced distinct LDL bands with identical migration distances when stained with Sudan black B or with Coomassie Brilliant blue (Fig. 1). Per serum sample analyzed the migration distance of LDL visualized by Sudan black B and by Coomassie Brilliant blue differed by an average ( f SD) of 0.2 k 2.1% in 32 data pairs. Of sixteen patients with a total of 32 serum samples, 29 samples (91%) revealed more than one LDL band after electrophoresis. Three serum samples (9%) had LDL electrophoretic pattern showing one band only. We considered the particle size of the major LDL band to be the representative particle size of LDL in the present study. The relationship between the particle size of the major LDL band and the concentrations of lipids and LDL-Apo B in serum was analyzed in the 32 serum samples obtained (Table I). During Dsimvastatin ther$py LDL particle size increased significantly (from 233 + 5.0 A to 237 k 7.0 A, P < 0.05). Meanwhile, the concentrations of total cholesterol, total triglyceride, VLDL-cholesterol, VLDL-triglyceride, LDLcholesterol, and LDL-Apo B in serum decreased significantly, whereas the concentration of HDL-cholesterol in serum increased significantly, but the LDLcholesterol to LDL-Apo B ratio did not change (Table II). The increase of LDL particle size during simvastatin therapy correlated significantly to the decreases of the concentrations of total triglyceride (r = -0.557, P < 0.05), VLDL-cholesterol (r = - 0.592, P < O.OS>, VLDL-triglyceride (r = -0.662, P < O.Ol), and to the increase of the concentration of HDL-cholesterol (r = 0.514, P < 0.05) in serum (Table III, Fig. 2). Stepwise multiple regression analysis of therapy-induced changes demonstrated that the decrease of VLDL-triglyceride concentration in serum correlated best (r2 = 0.438) with the increase of LDL particle size. The best two variables were the decrease of VLDL-triglyceride concentration, and the increase of HDL-cholesterol concentration (r2 = 0.596) in
TABLE I Correlations of LDL particle size with the concentrations samples studied
of lipids and LDL-Apo B in serum in 32
Parameters
Correlation coefficient
P
Total cholesterol Total triglyceride VLDL-cholesterol VLDL-triglyceride HDL-cholesterol LDL-cholesterol LDL-Apo B LDL-cholesterol/Ape
0.158 - 0.694 - 0.595 - 0.735 0.572 0.248 0.095 0.453
n.s. < 0.001 < 0.001 < 0.001 < 0.001 n.s. ns. < 0.01
IIS., Not significant.
B
114 TABLE II Mean values (*SD) and relative changes of LDL particle size, the concentrations LDL-Apo B in serum before and during simvastatin therapy (N = 16)
of lipids and
Parameters
Before therapy
During therapy
Changes (%I
P **
LDL particle size (A) Total cholesterol (mmol/l) Total triglyceride (mmol/l) VLDL-choiesterot (mmolfl) VLD~triglyce~de (mmol/l) HDL-cholesterol (mmol/l) LDL-cholesterol (mmol/l) LDL-Apo B (mg/dl) LDL-cholesterol/Ape B
233 * 5.01 11.2 * 3.05 3.78i 0.92 1.80& 0.72 2.67+ 0.87 0.8Ok 0.16 8.76+ 3.27 173 *49.7 1.955 0.32
237 f 7.65 f 2.56* 0.92* 1.80-f 0.91 f 5.85* 125 f 1.8Ok
1.8 * -30.1 * - 29.8 * -43.1 * -28.0 * 13.5 * -35.7 * -26.3 * - 5.5
< < < < < < <
109
0 1991 Elsevier Science Publishers B.V. All rights reserved 0009-8981{91/$03.50
CCA 05124
Effect of simvastatin on the apparent size of LDL particles in patients with type IIB hyperlipoproteinemia Shui P. Zhao ‘, Leny Hollaar Augustinus
‘, Ferdinand M. van ‘t Hooft H.M. Smelt *, Jan A. Gevers Leuven 3
‘,
and Arnbud van der Laarse ’ Departments of ’ Cardiology and ’ Internal Medicine, University Hospital, Leiden, and ’ Gaubius Institute TNO, Leiden (The Netherlands)
(Received 17 July 1990; revision received 19 August 1991; accepted 21 August 1991) Key words: Low density lipoprotein; LDL particle size; Type IIb hyperlipoproteinemia;
Simvastatin
Summary After 15 weeks of simvastatin therapy (20 mg/day), low density lipoprotein particle size in sera of 16 patjents with type I,‘b hyperlipoproteinemia increased significantly from 233 f 5.0 A to 237 f 7.0 A (P < 0.051, analyzed by 2-16% polyacrylamide gradient gel electrophoresis. Under simvastatin therapy the concentrations of total cholesterol, total triglyceride, very low density lipoprotein cholesterol and triglyceride, low density lipoprotein cholesterol and apolipoprotein B in serum fell significantly by 30%, 30%, 43%, 28%, 36% and 26%, respectively, and the concentration of high density lipoprotein cholesterol rose significantly by 14%. The changes of low density lipoprotein particle size induced by simvastatin therapy were correlated best with the changes of very low density lipoprotein triglyceride concentration (r2 = 0.438, P < 0.01). Our results suggest that simvastatin therapy, additionally to a reduction of the serum cholesterol concentration, increases low density lipoprotein particle size which may contribute to reduction of the risk of coronary heart disease in patients with type IIb hyperlipoproteinemia. Introduction Low density lipoprotein (LDL) isolated from human plasma contains approximately 50% cholesterol (free and esterified), 25% protein, 20% phospholipid and
Correspondence to: Dr. A. van dei Laarse, Department of Cardiology, Building 1. CS-P24, University Hospital, Rijnsburgenveg 10, 2333 AA Leiden, The Netherlands.
110
5% triglyceride. The heterogeneity of LDL with respect to size and density has been demonstrated by several groups [l-5]. Using 2-16% polyacrylamide gradient gel electrophoresis, Krauss and Burke [2] have demonstrated the presence of seven distinct LDL bands. They showed that these bands of LDL are associated with different LDL densities by means of density gradient ultracentrifugation. Recently, McNamara et al. [l] found that LDL subfraction distribution is influenced by gender and plasma lipoprotein levels. The characteristics of LDL subfractions as described by them were comparable to those observed after gradient ultracentrifugation [3]. Furthermore, Crouse et al. [4] reported a higher frequency of small LDL in patients with premature coronary artery disease than in healthy individuals. A recent study on LDL subclass patterns supports this observation [5]. It is now agreed that the heterogeneity in physical (size and density), chemical, and immunological properties of LDL is an inherent characteristic of circulating LDL [1,5-71. Although it has been proposed that the distribution of LDL subfractions is determined genetically [7-91, several studies have demonstrated that the concentrations of lipids in serum, especially triglyceride, influence the migration distance and patterns of LDL bands on polyacrylamide gradient gel electrophoresis [1,2,5] and LDL subfraction distribution after gradient ultracentrifugation [3]. The purpose of the present study is to examine whether LDL particle size is affected by simvastatin therapy in patients with type IIb hyperlipoproteinemia and if so, whether therapy induced changes of LDL particle size are related to changes in serum lipid levels. Materials and methods Patients
Sixteen patients (11 male and 5 female) with type IIb hyperlipoproteinemia were studied from November 1988 to January 1990. The ages of the patients ranged from 37 to 66 years, with an average age of 53.4 yr. The diagnosis of type IIb hyperlipoproteinemia was based on a serum total cholesterol concentration > 8 mmol/l, serum total triglyceride concentration > 2 mmol/l, LDL-cholesterol concentration > 4 mmol/l, and very low density lipoprotein (VLDL) cholesterol concentration > 1 mmol/l. Thirteen patients had either a positive family history of hyperlipoproteinemia or premature development of atherosclerosis, and were considered to have familial combined hyperlipoproteinemia. Three patients had xanthoma with a positive family history of hyperlipoproteinemia or premature development of atherosclerosis, and were considered to have familial hypercholesterolemia. None of the selected patients was overweight, nor was any of them drinking alcohol (> l/2 ounce/day). None had renal, thyroid or liver disease, or diabetes. All patients were given a diet containing fat (30% of total calories) with a sat : mono : poly ratio of 1: 1: 1, and < 300 mg cholesterol per day for at least 6 wk. Then, the patients started to receive simvastatin (20 mg/day) therapy.
111
Methods In all patients venous blood samples were obtained before taking simvastatin and after 15 wk (average) of simvastatin therapy. Blood samples were taken in the morning after > 12 h of fasting. After clotting, serum was obtained by centrifugation at 1,344 x g for 10 min at 20°C < 4 h after sampling. Five ml of fresh serum was ultracentrifuged with 5 ml of normal saline in a Kontron 75.13 rotor at 412,000 X g for 15 h at 15°C (Kontron AG, Zurich, Switzerland). Supernatant contained VLDL, and infranatant contained LDL and high density lipoprotein (HDL). HDL-cholesterol was determined by precipitating LDL in the infranatant by using phosphotungstic acid and MgCI,. LDL-cholesterol was the cholesterol measured in the infranatant minus HDL-cholesterol. In serum and ultracentrifugal fractions the triglyceride and cholesterol concentrations were measured enzymatitally using test kits (Boehringer, Mannheim, FRG). LDL-apolipoprotein (APO) B in the ultracentrifugal infranatant was measured by immunonephelometry (Beckman, Brea, California, USA). The remaining serum was stored at -20°C. The two serum samples from each patient obtained before and during therapy were analyzed simultaneously in the same polyacrylamide gel. Each sample was analyzed in two gels, one stained with Sudan black B which interacts with fatty moieties of lipoprotein, and a second stained with Coomassie Brilliant blue which interacts with protein moieties of lipoprotein. The method of gradient gel electrophoresis was according to McNamara et al. [l]. Briefly, each lane of the 2-16% nondenaturating polyacrylamide gels (Pharmacia LKB, Uppsala, Sweden) ‘was loaded with 7.5 ~1 of mixed serum prepared by mixing three volumes of serum with one volume of a solution of 1.17 mol/l sucrose and 0.1% bromophenol blue. Pre-electrophoresis of samples at 70 V for 20 min at 8°C was followed by electrophoresis at 450 V for 6 h at 8°C. One gel was stained with Sudan black B for 15 h. The Sudan black B staining solution was prepared by adding 0.8 g of Sudan black B and 4.0 g of zinc acetate to 300 ml of 60% ethanol in distilled water and heating to boiling for 10 min with slow stirring. Then the staining solution was filtered hot using filter paper. After cooling to room temperature the solution was refiltered before use. The gel was destained with a 50% solution of Cellosolve (Sigma, St. Louis, MO, USA) in distilled water for 2 h. The second gel was stained with Coomassie Brilliant .blue for 2 h. This staining solution was prepared by adding 1.25 g of Coomassie Brilliant blue R-250 to 250 ml of methanol, and 300 ml of 2.91 mol/l acetic acid in distilled water. The gel was destained for 15 h with the same solution except for Coomassie Brilliant blue. After destaining, the gels were rehydrated in a solution of 25% Cellosolve in distilled water to restore gel size and shape. Stained gels were scanned on a LKB 2202 Ultrascan Laser Densitometer (LKB, Paramus, NJ, USA). In each gel, the reference standard serum obtained from a healthy volunteer was applied to lane 1, 6, and 11 of a total of 11 lanes. When gel length was 70 mm after rehydration (original length of gels ‘is 78-79 mm), the migration distance of reference standard LDL was 14 mm. Meanwhile, a set of standard proteins with known hydrated diameters (HMW electrophoresis calibration kit, Pharmacia,
112
Piscataway, NJ, USA) such as thyroglobulin (170 A>, ferritin (122 A>, and catalase (104 A> was run in a number of gels and stained with Coomassie Brilliant blue to calibrate the diameter of LDL particles.
Statistical analysis Results are expressed as means and standard deviations (mean f SD). Differences between mean values were statistically evaluated by paired Student’s z test. Correlation analysis and stepwise multiple regression analysis were performed using SPSS/PC +@ (SPSS Inc., Chicago, IL, USA) software.
1
I
Fig. 1. Gradient black B (bottom).
1
10
2
3
4
5
6
7
6
9
2
3
4
5
6
7
6
9
polyacrylamide Both gels were
slab gels stained with Coomassie loaded
identically
with normal
10
Brilliant
11
I
11
I
blue (top) or with Sudan
serum (lanes
1, 6 and II)
patient sera (lanes 2, 3, 4, 5, 7, 8, 9, IO). Sera were applied to top of each gel.
and with
113
Results Electrophoresis of serum on 2-16% polyacrylamide gradient gels produced distinct LDL bands with identical migration distances when stained with Sudan black B or with Coomassie Brilliant blue (Fig. 1). Per serum sample analyzed the migration distance of LDL visualized by Sudan black B and by Coomassie Brilliant blue differed by an average ( f SD) of 0.2 k 2.1% in 32 data pairs. Of sixteen patients with a total of 32 serum samples, 29 samples (91%) revealed more than one LDL band after electrophoresis. Three serum samples (9%) had LDL electrophoretic pattern showing one band only. We considered the particle size of the major LDL band to be the representative particle size of LDL in the present study. The relationship between the particle size of the major LDL band and the concentrations of lipids and LDL-Apo B in serum was analyzed in the 32 serum samples obtained (Table I). During Dsimvastatin ther$py LDL particle size increased significantly (from 233 + 5.0 A to 237 k 7.0 A, P < 0.05). Meanwhile, the concentrations of total cholesterol, total triglyceride, VLDL-cholesterol, VLDL-triglyceride, LDLcholesterol, and LDL-Apo B in serum decreased significantly, whereas the concentration of HDL-cholesterol in serum increased significantly, but the LDLcholesterol to LDL-Apo B ratio did not change (Table II). The increase of LDL particle size during simvastatin therapy correlated significantly to the decreases of the concentrations of total triglyceride (r = -0.557, P < 0.05), VLDL-cholesterol (r = - 0.592, P < O.OS>, VLDL-triglyceride (r = -0.662, P < O.Ol), and to the increase of the concentration of HDL-cholesterol (r = 0.514, P < 0.05) in serum (Table III, Fig. 2). Stepwise multiple regression analysis of therapy-induced changes demonstrated that the decrease of VLDL-triglyceride concentration in serum correlated best (r2 = 0.438) with the increase of LDL particle size. The best two variables were the decrease of VLDL-triglyceride concentration, and the increase of HDL-cholesterol concentration (r2 = 0.596) in
TABLE I Correlations of LDL particle size with the concentrations samples studied
of lipids and LDL-Apo B in serum in 32
Parameters
Correlation coefficient
P
Total cholesterol Total triglyceride VLDL-cholesterol VLDL-triglyceride HDL-cholesterol LDL-cholesterol LDL-Apo B LDL-cholesterol/Ape
0.158 - 0.694 - 0.595 - 0.735 0.572 0.248 0.095 0.453
n.s. < 0.001 < 0.001 < 0.001 < 0.001 n.s. ns. < 0.01
IIS., Not significant.
B
114 TABLE II Mean values (*SD) and relative changes of LDL particle size, the concentrations LDL-Apo B in serum before and during simvastatin therapy (N = 16)
of lipids and
Parameters
Before therapy
During therapy
Changes (%I
P **
LDL particle size (A) Total cholesterol (mmol/l) Total triglyceride (mmol/l) VLDL-choiesterot (mmolfl) VLD~triglyce~de (mmol/l) HDL-cholesterol (mmol/l) LDL-cholesterol (mmol/l) LDL-Apo B (mg/dl) LDL-cholesterol/Ape B
233 * 5.01 11.2 * 3.05 3.78i 0.92 1.80& 0.72 2.67+ 0.87 0.8Ok 0.16 8.76+ 3.27 173 *49.7 1.955 0.32
237 f 7.65 f 2.56* 0.92* 1.80-f 0.91 f 5.85* 125 f 1.8Ok
1.8 * -30.1 * - 29.8 * -43.1 * -28.0 * 13.5 * -35.7 * -26.3 * - 5.5
< < < < < < <
