Modeling the Pharmacokinetic Profile of

Prednisolone
Page Count: 20
Introduction
The escalating challenge of drug resistance isn't just a distant concern; I am personally affected

by the reality of drug resistance, which is an increasingly pressing issue. As a child, I personally

experienced my father's and grandmother's battles with arthritic conditions. Prednisolone and

other similar drugs were frequently used in their treatment to reduce the pain and inflammation

brought on by arthritis. Observing the effects of these drugs on their lives prompted a personal

curiosity about how drugs function in the body. This personal connection sparked my interest in

pharmacokinetics, the mathematical study of how drugs move through the body. Understanding

the complex mathematical dance of absorption, distribution, metabolism, and excretion that

determines a medication's effectiveness is more important than simply learning formulas. For

me, researching the pharmacokinetics of prednisolone is more than just academic pursuit; it's a

means of learning how to customize treatment plans for optimal outcomes. Through the

integration of my personal experiences with a more profound comprehension of

pharmacokinetics, I am motivated to make a significant contribution to the medical domain.

Applying mathematical concepts to clinical practice, in my opinion, can help us make better

decisions that fight drug resistance and enhance patient outcomes. Not only is this journey about

my professional goals, but it's also about actually improving the lives of people who are dealing

with chronic illnesses like arthritis.

1
Mathematical Background
The primary objective of this investigation is finding a suitable model that describes the plasma

concentration of prednisolone as a function of time is the main goal of this study.

Concentration-time graphs are widely used, providing medical professionals with important

insights into the field of pharmacokinetics (Scheff et al., 2011). Medical practitioners can gain

comprehensive understanding of the distinct properties of the medication by conducting a

comprehensive examination of the concentration-time curve. By examining concentration-time

graphs, pharmacokinetic parameters such as peak concentration, time to reach 𝐶𝑚𝑎𝑥 , and the area

under the curve is obtained. These parameters are essential for optimizing drug dosing and

predicting efficacy and safety. Below is a conventional concentration-time graph depicting the

pharmacokinetic profile of an administered drug. Additionally, an accompanying table elucidates

the distinctively labeled terminologies featured on the graph (Junaid et al., 2019).

Figure 1: Labeled Concentration-Time Graph

2
 Table 1: Definitions of Fundamental Pharmacokinetic Terminology
Term Definition

MEC (Minimum Effective Concentration) The minimal concentration of medication within the
bloodstream requisite for the pharmaceutical agent
to manifest its therapeutic efficacy.

MTC (Minimum Toxic Concentration): The minimal concentration of medication in the

bloodstream anticipated to elicit a toxic effect.

Therapeutic Range The interval, delimited by MTC and MEC, within

which the drug concentration achieves effectiveness
with negligible toxicity

Cmax (ng/ml) The maximum of plasma concentration attained by

the pharmaceutical substance.

tmax (h) The duration subsequent to administration at which

the drug reaches its peak plasma concentration.

AUC (h ng ml-1) An indicator reflecting the cumulative body

exposure to the drug subsequent to administration.

Methodology

In order to simulate a drug's concentration over time, this study used regression modeling and

polynomial interpolation as methodological approaches. Polynomial interpolation is a

mathematical method used to construct a polynomial function that passes through a given set of

data points. Given 𝑛 + 1 distinct bivariate data points (𝑥0, 𝑦0),(𝑥1, 𝑦1),...,(𝑥𝑛, 𝑦𝑛), where the 𝑥𝑛

values are distinct. The goal of polynomial interpolation is to find a polynomial function that fits

the given data points. A polynomial function is defined by coefficients 𝑎𝑛 which are established

by solving a system of linear equations formed by substituting the given data points into the

polynomial function (Hassoun). Here 𝑛 is the degree of the polynomial which is represented as:

2 𝑛
𝑃(𝑥) = 𝑎0 + 𝑎1𝑥 + 𝑎2𝑥 +... + 𝑎𝑛𝑥

3
Conversely, regression modeling includes techniques for both linear and nonlinear regression. A

linear equation is used in linear regression to model the relationship between a dependent

variable (e.g., outcome) and one or more independent variables (e.g., predictors) (Militkþ &

Meloun, 2011). A simple linear regression equation, Where 𝑦 is the dependent variable, 𝑥 is the

independent variable, β is the intercept, β1 is the slope, and ε is the error term, is denoted as:

𝑦 = β0 + β1𝑥 + ε

Nonlinear regression, on the other hand, uses a nonlinear function to characterize the relationship

between the dependent variable and the independent variable(s). This function can be any

equation that describes the underlying relationship in the data, including sigmoidal, polynomial,

exponential, logarithmic, power, and others. In both linear and nonlinear regression, parameter

estimation is an essential step (Militký & Meloun, 2011). It entails calculating the parameters of

the function that has been selected to fit the observed data the best. By adjusting the fitted curve's

shape, scale, and location, these parameters make it possible to find significant correlations

between the variables.

The investigation will then be carried out in-depth using differential and integral calculus

principles. In order to assess the accuracy of both models, this analytical endeavor seeks to

derive two separate pharmacokinetic profiles and clarify important parameters like 𝐶𝑚𝑎𝑥 and

𝑡𝑚𝑎𝑥. Nevertheless, a thorough evaluation of the body's exposure to prednisolone administration

cannot be obtained from these values alone. Therefore, it is essential to determine Prednisolone's

Area Under the Curve (AUC). Through the comparison of AUC values between the two models,

healthcare professionals can determine safe dosing regimens that will maximize therapeutic

effects while minimizing side effects and lower the risk of toxicity that comes with exceeding the

4
Maximum Tolerated Concentration (MTC) (Rolan & Griffine, 2009). As such, a formulation of

the pharmacokinetic profile of prednisolone that shows a lower MTC level is substantially safer.

Data Analysis

Appendix A describes how data from a clinical trial published in the Journal of

Pharmaceutics and Pharmacology is used to create a graphical representation. Prednisolone

concentration is plotted along the y-axis, and time is plotted as the independent variable along

the x-axis.

Figure 2: Scatter Graph of Time (h) vs. Serum Concentration of Prednisolone (ng/ml)

Time intervals from 𝑡 = 0 to 𝑡 = 4 are not as regular as those from 𝑡 = 4 to 𝑡 = 9, as Figure

2 shows. The crucial pharmacokinetic parameter, the peak serum concentration, which is

measured at 160.000 ng/ml, was especially selected to be tracked over the first three points in

5
time. After reaching Cmax., Concentration clearly starts to decrease, as can be seen by looking at

the graph. I chose to compute the average rate of change in order to obtain more insight into the

graph's behavior. To calculate the average rate of change or slope of the two intervals, apply the

slope formula as follows:

∆𝐶
𝑆𝑙𝑜𝑝𝑒 = ∆𝑡

1. Calculation for the slope between 𝑡 = 0 and 𝑡 = 4

(160−0)
𝑆𝑙𝑜𝑝𝑒 = (4−0)

𝑆𝑙𝑜𝑝𝑒 = 40

2. Calculation for the slope between 𝑡 = 4 and 𝑡 = 9

(64−160)
𝑆𝑙𝑜𝑝𝑒 = (9−4)

𝑆𝑙𝑜𝑝𝑒 =− 19. 2

By calculating the average rate over time intervals, we discern transitions in concavity. A shift

from positive to negative values indicates a shift from concave up to concave down. This

transition is visually evident in Figure 2: the graph ascends rapidly until 𝑡 = 4 hours, then

descends. This turning point signifies an inflection point. The data's non-linearity suggests

inadequacy of a linear model. Moreover, the asymmetry of Figure 1 undermines the applicability

of a quadratic model. Given likely singularity in the inflection point, a cubic model appears most

suitable.

6
Exploration
Method A: Polynomial Interpolation

The equation of a cubic function is denoted in the standard form:

3 2
𝑓(𝑥) = 𝑎𝑥 + 𝑏𝑥 + 𝑐𝑥 + 𝑑
The constant terms a, b, c, and d are all non-zero coefficients that correspond to different terms

3
in the polynomial, ranging from 𝑥 to the constant term. So, precisely four data points are needed

to define a cubic function uniquely. Four equations are produced, one for each coefficient, when

these points are entered into the function. An underdetermined system results from an inadequate

number of data points. On the other hand, a system is correctly determined when exactly four

points are present, guaranteeing the acquisition of a unique cubic function. The system becomes

overdetermined when more data points are added, which could result in the interpolation of the

best-fitting cubic function. I therefore need to collect 4 data points in order to build a cubic

polynomial model. I have deliberately selected four data points from the seven available to

ensure a fair distribution for this analysis. Consequently, in order to apply the binomial theorem

in this situation, one must ascertain how many ways there are to choose four data points from a

possible seven. The binomial theorem states that for any non-negative integers 𝑛 and 𝑘, where 𝑛

is the total number of items and 𝑘 is the number of items to choose, in this scenario, 𝑛 = 7 and

𝑘 = 4. Applying these variable in the binomial coefficient formula yields:

𝑘 𝑛!
𝐶𝑛 = 𝑘!(𝑛−𝑘)!

4 7!
𝐶7 = 4!(7−4)!

4
𝐶7 = 35

7
From a total of seven data points, there are 35 ways to choose four, each of which represents a

distinct subset for building a cubic polynomial model. The Lagrange interpolation formula offers

an alternative approach to finding the polynomial, avoiding the need to solve each of these 35

equations separately. By using this formula, a polynomial that precisely represents given values

of the independent variable at any point in time can be derived (Tatum, 2023).

The formula is denoted as:

𝑛
𝑃(𝑥) = ∑ 𝐿𝑖(𝑥)𝑦𝑖
𝑖=1

The notation 𝑃(𝑥) denotes the polynomial function under construction, aimed at precisely fitting

the given dataset. It is the function we seek, designed to seamlessly pass through each of the

provided data points. 𝑦𝑖 signifies the function values corresponding to each data point. Most

importantly 𝐿𝑖(𝑥), where 𝑖 ranges from 1 to 𝑛 Lagrange polynomials, refers to a set of 𝑛

Lagrange basis polynomials, each with a degree of 𝑛 − 1. These polynomials are pivotal in

interpolating the dataset, facilitating the construction of the polynomial function 𝑃(𝑥). They

ensure that the resulting polynomial seamlessly fits the given data points, thereby aiding in

accurate modeling and analysis (Tatum, 2023). These Lagrange basis polynomials are formulated

as:

𝑛 𝑥−𝑥1 (𝑥−𝑥1)...(𝑥−𝑥𝑖−1)(𝑥−𝑥𝑖+1)...(𝑥−𝑥𝑛)
𝐿𝑖(𝑥) = ∏ 𝑥𝑖−𝑥𝑗
= (𝑥𝑖−𝑥1)...(𝑥𝑖−𝑥𝑖−1)(𝑥𝑖−𝑥𝑖+1)...(𝑥𝑖−𝑥𝑛)
𝑗=1,𝑗≠𝑖

8
 𝑛
The symbol ∏ denotes a product operation, iterating over 𝑗 from 1 to 𝑛, except when 𝑗 is
𝑗=1,𝑗≠𝑖

equal to 𝑖. This operation involves the multiplication of 𝑛 − 1 terms in the context at hand.

𝑥−𝑥1
The fraction 𝑥𝑖−𝑥𝑗
represents the ratio of two terms:

● 𝑥 − 𝑥1: The numerator represents the difference in the x-coordinate between the ith and jth

data points, where j is any number between 1 and 𝑛, not including the case where j = i.

● 𝑥𝑖 − 𝑥𝑗: In the denominator delineates the difference between the x-coordinate at which

the Lagrange polynomial is being evaluated x and the x-coordinate of the 𝑗-th data point,

where 𝑗 ranges from 1 to 𝑛, excluding the scenario where 𝑗 is identical to 𝑖.

● The denominator defines the difference between the 𝑥-coordinate of the 𝑗th data point

(where 𝑗 is not equal to 𝑖) and the 𝑥-coordinate at which the Lagrange polynomial is

being evaluated (𝑥); this difference can be found for any 𝑗 between 1 and 𝑛.

Prednisolone concentration over time is modeled by a cubic equation that is easily found using

the Lagrange interpolation formula given the data points (0,0), (2, 130), (4, 160), and (6, 126),

which correspond to concentrations (C) at different times (t).

The initial step is to determine the 4 Lagrange polynomials.

𝑛 𝑡−𝑡1
𝐿𝑖(𝑡) = ∏ 𝑡𝑖−𝑡𝑗
𝑗=1,𝑗≠𝑖

9
 Therefore, the 4 Lagrange polynomials are:

(𝑡−2)(𝑡−4)(𝑡−6)
𝐿1(𝑡) = (0−2)(0−4)(0−6)

(𝑡−0)(𝑡−4)(𝑡−6)
𝐿2(𝑡) = (2−0)(2−4)(2−6)

(𝑡−0)(𝑡−2)(𝑡−6)
𝐿3(𝑡) = (4−0)(4−2)(4−6)

(𝑡−0)(𝑡−2)(𝑡−4)
𝐿4(𝑡) = (6−0)(6−2)(6−4)

These expressions must undergo simplification through the expansion of both the numerator and

the denominator for each of the four Lagrange polynomials, leading to the subsequent

formulations:
3 2
𝑡 −12𝑡 +44𝑡−48
𝐿1(𝑡) =− 48

3 2
𝑡 −10𝑡 +24𝑡
𝐿2(𝑡) = 16

3
𝑡 −18+12𝑡
𝐿3(𝑡) =− 16

3 2
𝑡 −6𝑡 +8𝑡
𝐿4(𝑡) = 48

Having derived each Lagrange polynomial for every given point, one can subsequently substitute

them into the formulation of a Lagrange interpolation..

4
𝑃(𝑡) = ∑ 𝐿𝑖(𝑡)𝐶𝑖
𝑖=1

𝑃(𝑡) = 𝐿1(𝑡)𝐶1 + 𝐿2(𝑡)𝐶2 + 𝐿3(𝑡)𝐶3 + 𝐿4(𝑡)𝐶4

3 2 3 2 3 3 2
𝑡 −12𝑡 +44𝑡−48 𝑡 −10𝑡 +24𝑡 𝑡 −18+12𝑡 𝑡 −6𝑡 +8𝑡
𝑃(𝑡) = (− 48
• 0) + ( 16
• 130) + (− 40
• 160) + ( 48
• 126)

3 3 2
𝑃(𝑡) = 4
𝑡 − 17𝑡 + 96𝑡 + 0

10
I subsequently plotted the model function utilizing the graphing software SciDAVis.

Figure 3: Graph illustrating the Cubic model and its AUC alongside the Original Data Points

The pharmacokinetic exposure parameters, Cmax. and tmax, are analyzed using the cubic

polynomial model by using differentiation and integration methods from calculus. Finding the

moment when prednisolone's serum concentration reaches its maximum is the goal. The cubic

model function must first be differentiated using the Power rule.

3 3 2
𝑃 ´( 𝑥 ) = 4
𝑡 − 17𝑡 + 96𝑡 + 0

9 2
𝑃 ´( 𝑥 ) = 4
𝑡 − 34𝑡

To identify critical points of 𝑃(𝑡), where we equate 𝑃 ´( 𝑥 )to zero, yielding:

9 2
4
𝑡 − 34𝑡 = 0

11
Solving this equation provides the values of 𝑡 at which critical points occur. These critical points

denote the maxima of 𝑃(𝑡). For the zeros of the resulting quadratic equation, one must simplify

the equation by multiplying each side of the equation by 4 in order to eliminate the fraction.

9 2
4
𝑡 − 34𝑡 = 0

2
9𝑡 − 136𝑡 = 0

Apply the simplified equation to the Quadratic formula:

2
−𝑏 ± 𝑏 −4𝑎𝑐
𝑡= 2𝑎

2
−(−136) ± (−136) −4(9)(0)
𝑡= 2(9)

136 ± 18496
𝑡= 18

136
𝑡 = 0 𝑜𝑟 9

However, the maximum serum concentration of the drug is observed post administration,

136
therefore we can reject 𝑡 = 0 and accept 𝑡 = 9
. Accordingly, the maximum value, 𝐶𝑚𝑎𝑥 of

the cubic polynomial model can be determined by evaluating the polynomial at the determined

136
value of 𝑡 = 9
.

136
𝐶𝑚𝑎𝑥 = 𝑃( 9
)

3 136 3 136 2 136

𝐶𝑚𝑎𝑥 = 4
• ( 9
) − 17 • ( 9
) + 96 • ( 9
)

𝐶𝑚𝑎𝑥 ≈ 156. 708 ng/ml

The maximum concentration attained by the cubic model closely approximates the 𝐶𝑚𝑎𝑥

observed in the data, specifically at 160.000 ng/ml. This alignment underscores a notable

strength of the cubic model in accurately capturing the drug concentration dynamics over time.

12
To ascertain the bioavailability of Prednisolone, one must compute the area under the curve

represented by the cubic polynomial model. In this case the AUC is determined by we integrate

the given cubic polynomial function with respect to 𝑡, from the lower limit of 0 to the upper limit

of 6. This process involves finding the antiderivative (indefinite integral) of the function: then

evaluating it at the upper and lower limits of integration, and taking the difference between these

values.

6
3 3 2
𝐴𝑈𝐶 = ∫( 4 𝑡 − 17𝑡 + 96𝑡 + 0)𝑑𝑡
0

6
3 4 17 3 96 2
𝐴𝑈𝐶 = ∫( 16 𝑡 − 3
𝑡 + 2
𝑡 + 𝐶)𝑑𝑡
0

Then evaluating the antiderivative at the upper and lower limits of integration, and taking the

difference between these values.

4 6
3 17 3 2
𝐴𝑈𝐶 = ⎡ 16 𝑡 − 𝑡 + 48𝑡 + 𝐶⎤
⎣ 3 ⎦0

3 4 17 3 96 2 3 3 17 3 96 2
𝐴𝑈𝐶 = ( 16 (6) − 3
(6) + 2
(6) + 𝐶) − ( 16 (0) − 3
(0) + 2
(0) + 𝐶)

𝐴𝑈𝐶 = 747. 000 h ng ml-1

Method B: Nonlinear Regression

Utilizing Curve Expert® software, which employs nonlinear least squares regression to

determine functions that optimize fitting to a provided dataset. Results are presented in order of

descending coefficient of determination (R2), indicating the degree of fit quality from highest to

lowest. A Hoerl function was deemed optimal for modeling the relationship between serum

13
concentration and time intervals. The Hoerl function demonstrates superior fit with an R² value

of 0.997.

The general equation of a Hoerl function is as follows:

𝑥 𝑐
𝑓(𝑥) = 𝑎𝑏 𝑥

The function's magnitude or amplitude is determined by the variable (𝑎), which also controls the

function's vertical scaling. If 𝑎 > 1, the function shows rising values as 𝑥 increases; if 𝑎 < 1,

dependent values decrease as 𝑥 increases. The variable (𝑏), the exponential function's base,

controls the rate at which the function grows or decays in relation to 𝑥. As 𝑥 grows, there is

exponential growth if 𝑏 > 1, and exponential decay if 𝑏 < 1. The position along the horizontal

axis where the function is evaluated is determined by the independent variable, (𝑥). The

function's shape is affected by x's exponent, (𝑐). A positive 𝑐 quickens the function's rate of

increase or decrease as 𝑥 advances, whereas a negative 𝑥 slows it down (Mikhaylov et al., 2021).

The parameters obtained from Curve Expert analysis, rounded to six decimal places, are as

follows:

𝑎 ≈ 121. 351956

𝑏 ≈ 0. 695165

𝑐 ≈ 1. 228443

Substituting the specified parameters into the general Hoerl function yields the pharmacokinetic

profile equation for Prednisolone:

𝑥 1.228443
𝑓(𝑥) = 121. 351956 • 0. 695165 𝑥

Following, I plotted the model function utilizing the graphing software SciDAVis.

14
 Figure 4: Graph illustrating the Hoerl model and its AUC alongside the Original Data Points

The assessment of the fit between the Hoerl nonlinear regression model and pharmacokinetic

exposure parameters, specifically Cmax and tmax, entails applying calculus-based differentiation

within the established analytical framework employed for the cubic model. To streamline the

process while ensuring accuracy, Desmos is utilized to graphically represent the derivative of the

Hoerl function and determine its roots.

15
 Figure 4: Graph of the Hoerl Function and the Derivative Function with its Roots

The peak serum concentration of the drug occurs after administration, indicating that the initial

time point, 𝑡 = 0, can be disregarded in favor of a later time point, 𝑡 = 3. 387. Consequently,

the nonlinear Hoerl regression model's maximum value, Cmax , can be obtained by evaluating the

Hoerl function at 𝑡 = 3. 387

𝐶𝑚𝑎𝑥 = 𝑓(3. 387)

(3.387) 1.228443
𝐶𝑚𝑎𝑥 = 121. 351956 • 0. 695165 (3. 387)

𝐶𝑚𝑎𝑥 ≈ 159. 812 ng/ml

To enhance the comparative analysis between the two types of models, the AUC. The AUC is

computed through integration to quantify the cumulative exposure of prednisolone to the body

over time. However, for the purpose of conciseness and practicality, the following definite

integral with a lower limit of 0 and upper limit bound of 9 will be evaluated using a TI-84 Plus

CE graphing calculator.

16
 6
𝑥 1.552169
𝐴𝑈𝐶 = ∫(105. 625000 • 0. 647834 𝑥 )𝑑𝑡
0

𝐴𝑈𝐶 ≈ 746. 759 h ng ml-1

Conclusion

The close proximity of the 𝐶𝑚𝑎𝑥 values in my two models correspond to a similar closeness in

their respective AUC values. This surprising similarity in accuracy between the models contrasts

with my initial expectations. In previous research on pharmacokinetic methods utilizing

polynomial interpolation, there was a consistent pattern of 𝐶(𝑡)→ ∞ as 𝑡→ ∞. This contradicts

fundamental pharmacokinetic principles, where concentrations typically peak before declining

asymptotically. Additionally, the perpetual increase in prednisolone concentration contradicts

biological plausibility due to metabolism and excretion (Garza, 2023). This anticipated pattern

did not, however, emerge during my investigation. Conversely, the pharmacokinetic profile of

my cubic model, which had a horizontal asymptote at 𝐶 = 0, was typical. The observed results

do not match my hypothesis, which can be explained by the possibility that different

interpolation techniques produce non-identical behaviors. When time tends to infinity, for

example, polynomial interpolation based on techniques such as Vandermonde matrices or

Newton's divided differences may not converge to 0. (Werner, 1984). Additionally, different

behaviors, such as divergence or convergence to a nonzero value, can be caused by the

interpolation points chosen and the particular form of the interpolating polynomial. Since both

models produced 𝐶𝑚𝑎𝑥 and AUC values that were surprisingly close to each other, a comparison

study to determine which modeling technique is better requires calculating the percentage error

for each model. This shall be accomplished using the following formula:

|𝑃𝑟𝑒𝑑𝑖𝑐𝑡𝑒𝑑 𝐶𝑜𝑛𝑐𝑒𝑛𝑡𝑟𝑎𝑡𝑖𝑜𝑛 − 𝑂𝑏𝑠𝑒𝑟𝑣𝑒𝑑 𝐶𝑜𝑛𝑐𝑒𝑛𝑡𝑟𝑎𝑡𝑖𝑜𝑛|

% 𝐸𝑟𝑟𝑜𝑟 = 𝑂𝑏𝑠𝑒𝑟𝑣𝑒𝑑 𝐶𝑜𝑛𝑐𝑒𝑛𝑡𝑟𝑎𝑡𝑖𝑜𝑛
• 100%

17
Table 2: Presents a comparison between the predicted and observed concentrations for the Cubic
and Hoerl models, including the Percentage Error
Serum
Time (h) Concentration of Cubic Model % Error Hoerl Model % Error
Prednisolone(ng/ml)

0.000 0.000 0.000 0.000 0.000 0.000

0.5.000 48.000 43.844 8.658 43.181 10.040

2.000 130.000 130.000 0.000 137.411 5.701

3.000 143.000 155.250 8.556 157.191 9.924

4.000 160.000 160.000 0.000 155.595 2.753

6.000 126.000 126.000 0.000 123.734 1.798

9.000 64.000 33.750 47.266 68.402 6.878

Contrary to prior assertions, it is apparent that while the cubic model may initially appear to

adhere to the general structure of the pharmacokinetic profile, a reassessment in light of the

comparative standard error between both models refutes this assumption. Examination of the

percent error values at each data point for both models reveals that the cubic model exhibits a

superior fit, aligning closely with the secondary data, while the Hoerl model fails to accurately

predict data points except at 𝑡 = 0. However, the aggregate percent error for each model

indicates a higher overall percent error for the cubic model compared to the Hoerl model. As a

result, the Hoerl model is the optimal option for illustrating the pharmacokinetic profile of

prednisolone because a lower total percent error in the model indicates increased accuracy and

consequently lowers the risk for pharmacists who use it to calculate the Area Under the Curve

(AUC), which measures the total exposure of prednisolone in the bloodstream by minimizing the

possibility of exceeding the Maximum Tolerated Concentration (MTC) and the related side

effects, such as osteoporosis and autoimmune diseases (Puckett, 2023).

18
Evaluation

Although the goal of this investigation was accomplished, it is crucial to acknowledge both its

merits and drawbacks. The analysis's balanced use of secondary data in the Hoerl and Cubic

Models is one of its most noteworthy strengths. This could be mathematically illustrated by

comparing the variance or standard deviation of the data points used in each model, which would

show how evenly distributed the data points were. Nevertheless, the cubic polynomial model's

dependence on just four of the seven available data points is a drawback. The percentage of data

points used in the model relative to all available data points can be used to mathematically

interpret this. For instance, the cubic model only includes about 57% of the available data if it

uses four of the seven data points, which could result in a higher percentage error. By utilizing all

seven data points, the Hoerl Model, in contrast, maximizes the amount of information integrated

into the model. By comparing the percentage of data points used in the Hoerl Model to all

available data points, which would be 100%, this can be mathematically represented. One more

restriction that needs to be noted is that both models rely on a single Prednisolone dosage. This

restriction could be overcome mathematically by creating models that take different Prednisolone

dosages and intervals into consideration. This could entail adding more variables to the

models—like dosage quantity or frequency—and modifying the model equations appropriately.

The formula below can be used to calculate concentration following a second dose:

−𝑘𝑡
𝐶(𝑡) = [𝐶1] + [𝐶1]𝑒

Where C(t) represents the concentration of Prednisolone at time (t). [𝐶1], is the initial

−𝑘𝑡
concentration, and 𝑘 is the elimination rate constant. The term 𝑒 denotes the decay of

19
 concentration over time due to elimination processes (Durgadevi, 2018). Furthermore,

broadening this investigation to simulate Prednisolone's pharmacokinetics at various dosages

may provide insightful information. Further insight may also be obtained by examining its

concentration-time profile when given intravenously as opposed to orally. In conclusion,

developing pharmacokinetic profiles through the use of integration techniques and polynomial

interpolation demonstrates the multidisciplinary application of mathematics in biology. Models

based on secondary data have been developed by analyzing pharmacokinetics using functions,

calculus, and interpolation techniques. The treatment of arthritic conditions is clarified by this

mathematical viewpoint, which may be of future concern to me and many other family members.

It emphasizes how mathematics can be used to better understand biochemical processes, which

enhances knowledge and ignites interest in medical research.

References
Durgadevi. (2018). Using differential equations to model drug concentrations.
https://www.pramanaresearch.org/gallery/prj-s22.pdf

20
Garza, A. Z. (2023a, July 4). Drug elimination. StatPearls [Internet].
https://www.ncbi.nlm.nih.gov/books/NBK547662/#:~:text=Drug%20elimination%20is%2
0the%20sum,encompass%20both%20metabolism%20and%20excretion.

Garza, A. Z. (2023b, July 4). Drug elimination. StatPearls [Internet].

https://www.ncbi.nlm.nih.gov/books/NBK547662/#:~:text=Drug%20elimination%20is%2
0the%20sum,encompass%20both%20metabolism%20and%20excretion.

Hassoun, M. (n.d.). Polynomial Interpolation.

https://neuron.eng.wayne.edu/auth/ece512/lecture17.pdf

Junaid, T., Wu, X., & Thanukrishnan, H. (2019, January 11). Therapeutic Drug Monitoring.
Clinical Pharmacy Education, Practice and Research.
https://www.sciencedirect.com/science/article/abs/pii/B9780128142769000301#:~:text=Th
e%20minimum%20effective%20concentration%20(MEC,drug%20produces%20unwanted
%20side%20effects.

Mikhaylov, P. V., Shevelev, S. L., Sul’tson, S. M., Verkhovets, S. V., & Goroshko, A. A. (2021).
Modeling of growth and development of spruce stands in the west siberian south taiga
plain forest region. IOP Conference Series: Earth and Environmental Science, 677(5).
https://doi.org/10.1088/1755-1315/677/5/052121

Militký, J., & Meloun, M. (2011). Nonlinear regression model. Science Direct.
https://www.sciencedirect.com/topics/mathematics/nonlinear-regression-model

Puckett, Y. (2023, July 19). Prednisolone. StatPearls [.

https://www.ncbi.nlm.nih.gov/books/NBK534809/

Rolan, P., & Griffine, J. P. (2009). Chapter 5: Clinical Pharmacokinetics. In The Textbook of
Pharmaceutical Medicine, 6th Edition (6th ed.). essay, Blackwell Ltd.

Scheff, J. D., Almon, R. R., Dubois, D. C., Jusko, W. J., & Androulakis, I. P. (2011, May).
Assessment of pharmacologic area under the curve when baselines are variable.
Pharmaceutical research. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3152796/

Tatum, J. (2023, December 28). 1.11: Fitting a polynomial to a set of points - Lagrange
polynomials and Lagrange interpolation. Physics LibreTexts.
https://phys.libretexts.org/Bookshelves/Astronomy__Cosmology/Celestial_Mechanics_(Ta
tum)/01%3A_Numerical_Methods/1.11%3A_Fitting_a_Polynomial_to_a_Set_of_Points_-
_Lagrange_Polynomials_and_Lagrange_Interpolation

Werner, W. (1984). Polynomial interpolation: Lagrange versus Newton. Mathematics of

Computation, 43(167), 205. https://doi.org/10.2307/2007406

21
 Appendix A
Secondary Data of Prednisolone

22
I sourced the data for my investigation from a scholarly article titled "A Mathematical Model of

the Pharmacokinetic Dynamics of Orally Administered Prednisolone in Healthy Volunteers,"

published by the Journal of Pharmaceutics and Pharmacology.

Table A: Serum Concentration of Prednisolone over Time

Time (h) Serum Concentration of Prednisolone (ng/ml)

0.000 0.000

0.500 48.000

2.000 130.00

3.000 143.000

4.000 160.000

6.000 126.000

9.000 64.000

23