p r i m a r y c a r e d i a b e t e s 1 2 ( 2 0 1 8 ) 184–191

Contents lists available at ScienceDirect

Primary Care Diabetes

journal homepage: http://www.elsevier.com/locate/pcd

Original research

Prescription patterns and costs of antidiabetic

medications in a large group of patients

Andrés Gaviria-Mendoza, Jorge Andrés Sánchez-Duque,

Diego Alejandro Medina-Morales, Jorge Enrique Machado-Alba ∗
Grupo de Investigación en Farmacoepidemiología y Farmacovigilancia, Universidad Tecnológica de
Pereira-Audifarma S.A., Pereira, Colombia

a r t i c l e i n f o a b s t r a c t

Article history: Aims: To determine the prescription patterns of antidiabetic medications and the variables
Received 20 April 2017 associated with their use in a Colombian population.
Received in revised form Methods: A cross-sectional study using a systematized database of approximately 3.5 mil-
1 November 2017 lion affiliates of the Colombian Health System. Patients of both genders and all ages
Accepted 10 November 2017 treated uninterruptedly with antidiabetic medications for three months (June–August 2015)
Available online 28 November 2017 were included. A database was designed that included sociodemographic, pharmacological,
comedication, and cost variables.
Keywords: Results: A total of 47,532 patients were identified; the mean age was 65.5 years, and 56.3%
Diabetes mellitus were women. Among the patients, 56.2% (n = 26,691) received medication as monother-
Drug costs apy. The most prescribed medications were metformin, 81.3% (n = 38,664), insulins, 33.3%
Drug prescriptions (n = 15,848), and sulfonylureas, 21.8% (n = 10,370). Among the patients, 92.8% received
Hipoglycemic agents comedications, including antihypertensives (79.7%), hypolipemiants (65.5%), antiplatelet
Pharmacoepidemiology drugs (56.3%), analgesics (33.9%), antiulcerants (33.1%), and thyroid hormone (17.3%). The
Economics cost per 1000 inhabitants/day was $1.21 USD for metformin, $3.89 USD for insulins, and
Pharmaceutical $0.02 USD for glibenclamide.
Conclusions: Generally, rational prescription habits predominated, however in some cases an
overuse of comedications (such as antiulcer drugs) and a large group of patients with high
cost formulations were observed. Subsequent effectiveness and cost-benefit analyzes are
required.
© 2017 Primary Care Diabetes Europe. Published by Elsevier Ltd. All rights reserved.

were an estimated 381 million people with DM worldwide.

1. Introduction The majority resided in middle and low income countries.
However, by the year 2035, this figure may reach 592 million
Diabetes mellitus (DM) is a public health problem that affects
patients, corresponding to an increase of 55% [1,2]. According
approximately 8% of the global population. In 2013, there

∗
Corresponding author at: Calle 105 No. 14-140. Pereira, Risaralda 660003, Colombia.
E-mail address: machado@utp.edu.co (J.E. Machado-Alba).
https://doi.org/10.1016/j.pcd.2017.11.002
1751-9918/© 2017 Primary Care Diabetes Europe. Published by Elsevier Ltd. All rights reserved.
 p r i m a r y c a r e d i a b e t e s 1 2 ( 2 0 1 8 ) 184–191
to data from the World Health Organization, type 2 diabetes
mellitus corresponds to 90–95% of the cases of the disease.
In Latin America, the prevalence varies considerably between
countries, from 4.3% in Peru to 15.4% in Puerto Rico. In
Argentina, Chile, and Mexico, the percentages are 6.0%, 10.3%,
and 11.7%, respectively [3].
After Brazil and Mexico, Colombia has the third largest
population in Latin America, with approximately 48 million
inhabitants. The prevalence of type 2 diabetes mellitus in the
country is between 1.2% and 11.2%, whereas type 1 diabetes
mellitus is present in 0.07% of the population [1,4–6].
Currently, the pharmacotherapeutic options have signif-
icantly expanded due to the large number of medications.
Human insulin is the medication of choice for type 1 dia-
betes mellitus and advanced stages of type 2 diabetes
mellitus, whereas in other patients, metformin is the most
highly valued medication because of its ability to control
glycemic levels and to offer additional effects with sig-
nificant clinical improvement [7,8]. Other drugs used for
the management of DM are the sulfonylureas, thiazolidine-
diones, glucagon-like peptide-1 (GLP-1) analogs, dipeptidyl
peptidase-4 (DPP-4) inhibitors, sodium-glucose co-transporter
2 (SGLT-2) inhibitors, meglitinides, and alpha-glucosidase
inhibitors [9,10].
The Health System of Colombia provides universal cover-
age through two regimes: the first is paid by the user, and the
second is subsidized by the state. However, both have a bene-
fits plan known as the Mandatory Health Plan (Plan Obligatorio
de Salud, POS, in Spanish), which covers most drugs for the
treatment of DM, including all insulins and some formulations
of metformin and sulfonylureas, but does not yet include the
other molecules or combination drugs. To access the latter,
the treating health professional can request the medication
through a mechanism called the Scientific Technical Commit-
tee (Comité Técnico Científico, in Spanish) or through a legal
guardianship tool (“tutela”).
In Colombia, all patients with diabetes are treated routinely
by general practitioners, but are seen by internal medicine
and endocrinology depending on the clinical scenario (such
as major diabetes complications or those patients in whom
the goal is difficult to achieve). There is not legal restriction
for physicians to prescribe any of the antidiabetic medications,
but higher cost insulins or dosage forms are commonly started
by the specialist. General practitioners usually continue pre-
scribing the medications started by specialists.
The objective of this study was to determine the antidi-
abetic medication prescription patterns in a population of
people affiliated with the Colombian Health System.
2. Materials and methods
2.1. Study design and participants
A descriptive cross-sectional study of antidiabetic medica-
tion prescription behavior was conducted in a population of
approximately 3.5 million people affiliated with the contrib-
utory regime of the Colombian Health System in five Health
Promotion Entities (Insurance companies), which corresponds
to approximately 17.5% of the population actively affiliated
185
with this regime in the country and 7.3% of the Colombian
population.
Data from patients with dispensation of antidiabetic med-
ications between June 1 and August 31, 2015, of all ages and
either gender whose treatment was maintained continuously
for at least 3 months were included. This requirement was
set to ensure the inclusion of patients with stable and con-
tinuous treatment adherence because these are chronic use
medications.
From the information on the consumption of medica-
tions systematically obtained by the dispensing company
(Audifarma S.A.), a database was designed and collected the
following groups of variables:
1. Sociodemographic variables: age, gender, and city.
2. Antidiabetic medications available: biguanides (met-
formin), sulfonylureas (glibenclamide, gliclazide, and
glimepiride), thiazolidinediones (pioglitazone), GLP-1
analogs (exenatide and liraglutide), DDP-4 inhibitors
(sitagliptin, vildagliptin, saxagliptin, and linagliptin),
insulins (ultrashort action: lispro, aspartat, and glulisine;
short: regular; intermediate: NPH; long: glargine, detemir,
and degludec), meglitinides (nateglinide and repaglin-
ide), alpha-glucosidase inhibitors (acarbose), and SGLT-2
inhibitors (dapagliflozin and empagliflozin). Information
on the dose used was analyzed; the unit considered was
the defined daily dose (DDD) and its estimation per 1000
inhabitants/day (DID).
To establish comorbidity, comedication was accepted
as a surrogate indicator of chronic disease as follows:
(a) analgesics/pain management; (b) psychoneural drugs
(anxiolytics and hypnotics, antidepressants, antipsychotics,
and antiepileptics)/neurological and psychiatric disorders;
(c) antiplatelet drugs/cardiovascular prevention; (d) antiar-
rhythmics/cardiac arrhythmia; (e) anticoagulants/atrial fib-
rillation and venous thrombosis; (f) anti-hypertensives
and diuretics/arterial hypertension; (g) antiulcerants/acid-
peptic disease; (h) bisphosphonates/osteoporosis; (i) inhaled
bronchodilators/chronic obstructive pulmonary disease or
asthma; (j) estrogens and progestogens/contraception or
hormone replacement therapy; (k) gabapentin and pre-
gabalin/neuropathic pain; (l) hypolipemiants/dyslipidemia;
(m) thyroid hormone and antithyroids/thyroid disorders; (n)
inotropes/cardiac failure; and (o) nitrovasodilators/ischemic
heart disease.
For the evaluation of the economic impact of the use of
antidiabetic drugs in the Colombian Health System, the refer-
ence prices of the formulations most commonly used by the
different insurance companies were used. Then, the cost per
1000 inhabitants/day (CID) was evaluated (cost/[365 × number
of inhabitants] × 1000). Additionally, the monthly and annual
costs were estimated (reference value according to the Bank
of the Republic of Colombia, $1 USD = 3101 COP [August 31,
2015]).
The protocol received the approval of the Bioethics Com-
mittee of the Universidad Tecnológica de Pereira in the
category of “research without risk”. The protocol safeguarded
186 p r i m a r y c a r e d i a b e t e s 1 2 ( 2 0 1 8 ) 184–191
Fig. 1 – Distribution of 47532 patients treated with
antidiabetics, by age and sex, Colombia, 2015.
the identities of the patients and respected the principles
established by the Declaration of Helsinki.
2.2. Statistical analysis
® ® ®
IBM SPSS Statistics software, version 23.0 (IBM , USA)
®
for Windows , was used for the data analysis. Descriptive
statistics were employed. Student’s t-test and ANOVA were
performed to compare the quantitative variables, and X2 was
used to compare the categorical variables.
Non-parsimonious binary logistic regression models were
performed with the use of antidiabetic drugs alone or asso-
ciated (yes/no) and the need for comedication (yes/no) as
the dependent variables. The model was run with an entry
method and covariates included age category, sex, and those
variables significantly associated with the dependent vari-
ables in the bivariate analysis. A correlation matrix was run
for each analysis to check multicollinearity and interaction
terms were also assessed. A value of p < 0.05 was determined
as the level of statistical significance.
3. Results
A total of 47,532 patients who received antidiabetic treatment
during the study period were included. Of these, 26,764 (56.3%)
were women, and 20,768 (43.7%) were men. The mean age
was 65.5 ± 12.8 years (range: 4–106 years), with the following
distribution: <20 years (n = 238, 0.5%); 20–44 years (n = 2192,
4.6%); 45–64 years (n = 16,567, 34.9%); and ≥65 years (n = 22,630,
47.6%). Fig. 1 shows the distribution of this group of patients by
age and gender. No differences by ethnicity were established.
All medications and formulations available in Colombia for
the management of DM were considered. The insulin prescrip-
tion patterns are described in Table 1, and the prescription
patterns of the other antidiabetic medications are shown in
Table 2.
Fig. 2 – Frequency of prescription of main antidiabetics
alone or in combination therapy, Colombia, 2015.
3.1. Monotherapy versus combination therapy
Of the total patients in the study, 26,691 (56.2%) were treated
with a single antidiabetic medication, and 20,841 (43.8%)
received an association of antidiabetic medications, of which
17,431 (36.7%) had two prescribed, 3286 (6.9%) had three, and
124 patients (0.3%) had four or more.
Among the patients, 15,848 (33.3%) were indicated for
insulin therapy. Of these patients, 10,007 (21.1%) used insulin
in combination with another antidiabetic. Of the total peo-
ple included, 8524 (17.9%) used a single type of insulin and
7324 (15.4%) used more than one type of insulin. The distribu-
tion by frequency (from highest to lowest prescription) for the
long-acting, ultrashort-acting, intermediate, and short-acting
medications was 24.1%, 12.8%, 10.1%, and 2.7%, respectively.
A total of 11,462 people (24.1%, equivalent to 72.3% of those
with any insulin) used insulin analogs, and 4937 (10.4%) used
human insulins.
Regarding oral antidiabetic drugs, metformin was the most
used medication, being prescribed in 38,664 patients (81.3%)
who used some of its formulations. Of these, 15,731 (33.1%)
used it with another antidiabetic medication. Sulfonylureas
were the second most prescribed group of oral medications
(21.8%, n = 10,370), followed by DPP-4 inhibitors (13.7%, n = 6510
patients). GLP-1 analogs were prescribed to 419 (0.9%) patients,
of which 359 employed them as a comedication. Only 24
patients received a SGLT-2 inhibitor (23 of whom did so
together with another antidiabetic), and four patients received
acarbose (three as comedication). No patients were found with
prescriptions for thiazolidinediones and meglitinides.
3.2. Combination therapy
Fig. 2 shows the monotherapy/combined therapy relationship
for the evaluated antidiabetic medications. The use of met-
formin as a monotherapy and combination therapy for the
other antidiabetic medications is highlighted.
Among the 20,843 (43.9%) patients who were prescribed
antidiabetic combination therapy, the most frequently used
were insulin plus metformin (n = 9030 patients, 19.0%) and
metformin plus glibenclamide (n = 8629, 18.2%).
When the relationship between the use of combined ther-
apy and the other variables was analyzed by binary logistic
regression, we found that male sex, being prescribed neuro-
pathic pain medications, antiplatelet drugs, having ≥45 years
 p r i m a r y c a r e d i a b e t e s 1 2 ( 2 0 1 8 ) 184–191 187

Table 1 – Prescribing patterns of insulins in a population of 47532 patients from Colombia, 2015.
Insulin Prescriptions/users Prescribed doses DDD Female:male Mean age
(Units/day) ratio (SD)

# Patients % Mean Mode

Glulisine
Pen 100 U/mL/3 mL 2785 5.9 25.3 20 0.74 1.3:1 60.7 (16.7)
Vial 100 U/mL/10 mL 857 1.8 40.5 33 1.4:1 64.5 (14.3)
Cartridge 100 U/mL/3 mL 152 0.3 14.0 10 1.0:1 61.4 (20.5)

Aspartat pen 100 U/mL/3 mL 1522 3.2 27.6 20 0.69 1.4:1 62.27 (15.8)

Lispro
Pen 100 U/mL/3 mL 625 1.3 24.4 20 0.82 1.2:1 55.9 (19.9)
Vial 100 U/mL/10 mL 356 0.7 42.6 33 1.5:1 58.2 (18.1)
Cartridge 100 U/mL/3 mL 70 0.1 21.3 10 0.9:1 44.8 (26.9)
Cartridge 100 U/mL/3 mLa 13 0.0 67.7 60 0.9:1 59.4 (14.2)
Pen 100 U/mL/3 mLa 2 0.0 50.0 – 1.0:1 74.3 (9.1)

Regular vial 100 U/mL/10 mL 1260 2.7 32.4 33 0.81 1.4:1 66.7 (11.3)

NPH
Vial 100 U/mL/10 mL 4593 9.7 42.8 33 1.06 1.3:1 66.5 (11.9)
Cartridge 100 U/mL/3 mL 60 0.1 14.0 3 0.9:1 58.5 (14.9)
Pen 100 U/mL/3 mL 44 0.1 19.4 20 1.1:1 61.0 (11.3)
NPH + regular vial 70 + 30 U/mL/10 mL 9 0.0 57.9/24.81 31 0.8:1 71.9 (11.9)
NPH + regular cartridge 70 + 30 U/mL/3 mL 6 0.0 38.5/16.5 35 0.5:1 55.4 (27.8)

Glargine
Pen 100 U/mL/3 mL 6010 12.6 30.2 20 0.90 1.2:1 62.2 (15.9)
Vial 100 U/mL/10 mL 3120 6.6 44.9 33 1.3:1 64.7 (13.9)
Cartridge 100 U/mL/3 mL 112 0.2 14.4 3 1.0:1 59.8 (18.4)

Detemir Pen 100 U/mL/3 mL 2194 4.6 35.0 20 0.87 1.4:1 64.8 (13.7)
Degludec Pen 100 U/mL/3 mL 151 0.3 27.3 20 0.68 2.2:1 57.9 (17.6)
Insulin total (SD) 15848 33.3 52.1 33 1.30 (0.87) 1.3:1 64.2 (14.4)
DDD: average ratio between the prescribed daily dose and the defined daily dose; SD: standard deviation.
a
25% soluble — 75% protamine.

and being treated in the city of Palmira were significantly
associated with a higher likelihood of being treated with
antidiabetic combination therapy (Supplementary Table S1).
In contrast, receiving some comedications such as levothyrox-
ine or inhaled bronchodilators, and being treated in the cities
of Bogotá, Cali, Pereira, and Manizales were associated with
a lower probability of receiving several antidiabetic medica-
tions.
3.3. Comedication
Of the population included in the study, 44,106 (92.8%) received
concomitant treatment with one or more medications for
the most common comorbidities that accompany DM in
patients. Among the patients, we found 4258 with a single
co-medication (9.0% of cases), but patients with two (n = 8167,
17.2%), three (n = 12,403, 26.1%), four (n = 9640, 20.3%), and
up to five or more (n = 9638, 20.3%) additional drugs were
also found. In order of frequency, the most prescribed groups
of medications that may have potential interactions with
antidiabetic agents include anti-hypertensives (n = 37,873,
79.7%), hypolipemiants (n = 31,152, 65.5%; of which 90.4%
were statins), antiplatelet drugs (n = 26,737, 56.3%, 98.1%;
corresponding to 100 mg of acetylsalicylic acid), analgesics
(n = 16,133, 33.9%), antiulcerants (n = 15,716, 33.1%), levothy-
roxine (n = 8187, 17.2%), psycho and neurological medications
(n = 7068, 14.9%), inhaled bronchodilators (n = 2863, 6.0%), and
others (n = 2967, 6.26%).
Regarding antihypertensive medications, we determined
that the renin–angiotensin–aldosterone system (RAAS)
inhibitors were the most employed (n = 35,020, 92.5% of
patients with antihypertensives) and were distributed
between angiotensin-II receptor blockers (ARB-II) (n = 26,724,
70.6%) and angiotensin-converting enzyme (ACE) inhibitors
(n = 8,296, 21.9%), followed by calcium channel blockers
(n = 15,362, 40.6%), ␤-blockers (n = 12,284, 32.4%), thiazide
diuretics (n = 8513, 22.5%), and loop diuretics (n = 7446,
19.7%). Among the patients, 29.4% (n = 11,132) used only one
antihypertensive and 58.4% used two or three.
When the relationship between the use of comedication
and the other variables was analyzed using logistic regres-
sion, we found that using insulin, sulfonylureas, metformin,
or being 45 years or older increased the probability that the
patient received other groups of medications. In contrast, liv-
ing in the city of Bogotá, being male, or using GLP-1 analogs
was associated with a lower probability of receiving comedi-
cation (Supplementary Table S2).
We found some combinations of medications that required
special care because they could interfere with the proper
metabolic control of DM, enhance electrolyte disturbances,
or mask symptoms of hypoglycemia, such as diuretics
(n = 15,959, 33.6% of patients), alpha 2 adrenergic agonists
(1604, 3.4%), corticoids (1308, 2.8%), or even ␤-blockers.
188 p r i m a r y c a r e d i a b e t e s 1 2 ( 2 0 1 8 ) 184–191

Table 2 – Prescribing patterns of antidiabetics (excluding insulins) in a population of 47,532 patients from Colombia, 2015.
Medication Prescriptions/users Prescribed doses DDD Female:male Mean age
(mg/day) ratio (SD)

# Patients % Mean Mode

Metformin
Tab 850 mg 33,898 71.3 1478.4 850 0.75 1.3:1 65.7 (11.9)
Tab 1000 mg 417 0.9 1320.1 2000 1.1:1 61.8 (10.9)
Tab 500 mg 392 0.8 621.5 500 1.9:1 64.9 (12.3)
Tab 1000 mga 40 0.1 1325.0 1000 1.4:1 59.8 (12.1)
Tab 750 mga 23 0.0 1097.8 750 2.3:1 62.0 (11.6)
Tab 50 mga 21 0.0 627.0 500 1.6:1 61.6 (16.3)
Metformin + glibenclamide, tab 500 + 5 mg 21 0.0 849.21/8.49 1000 0.5:1 65.6 (7.8)
Metformin + glibenclamide, tab 500 + 2.5 mg 13 0.0 1025.64/5.13 1000 0.4:1 66.5 (6.9)
Tab 850 mga 4 0.0 850.0 566 1.0:1 68.4 (19.5)

Glibenclamide tab 5 mg 9895 20.8 8.5 5.0 0.85 1.1:1 65.9 (11.1)

Glimepiride
Tab 2 mg 188 0.4 2.2 2.0 1.83 1.2:1 66.3 (13.1)
Tab 4 mg 142 0.3 4.8 4.0 0.7:1 67.3 (10.1)
Glimepiride + metformin, tab 2 + 1000 mg 49 0.1 2.93/1467.57 3.9 1.2:1 64.1 (8.7)
Glimepiride + metformin, tab 4 + 1000 mg 36 0.1 6.56/1639.5 8.5 0.9:1 64.8 (8.8)
Glimepiride + metformin, tab 2 + 500 mg 25 0.1 3.2/800 4.0 0.8:1 69.1 (13.7)

Gliclazide
Tab 60 mga 66 0.1 67.57 60.0 1.17 1.4:1 63.4 (9.8)
Tab 80 mg 11 0.0 75.15 80.0 1.2:1 66.4 (8.2)

Liraglutide pen 6 mg/mL/3 mL 247 0.5 1.3 1.8 1.12 2.0:1 57.6 (10.5)

Exenatide
Powder for suspension 2 mg/3 mL 166 0.3 0.06 0.07 0.23 1.7:1 60.9 (10.3)
Pen Injection solution 250 mcg/mL/2.4 mL 4 0.0 0.02 0.02 0.3:1 60.5 (8.6)
Pen Injection solution 250 mcg/mL/1.2 mL 3 0.0 0.01 0.01 2.0:1 71.1 (9.1)

Sitagliptin
Sitagliptin + metformin, tab 50 + 1000 mg 1448 3.0 84.09/1681.73 93.33 0.84 1.2:1 64.4 (10.9)
Tab 100 mg 555 1.2 93.7 93.33 1.4:1 66.8 (11.4)
Tab 50 mg 319 0.7 64.7 46.66 1.5:1 67.8 (12.4)
Sitagliptin + metformin, tab 50 + 850 mg 315 0.7 77.96/1325.37 93.33 1.1:1 65.6 (11.3)
Sitagliptin + metformin, tab 50 + 500 mg 187 0.4 77.17/771.74 93.33 1.7:1 66.3 (11.0)

Vildagliptin
Vildagliptin + metformin, tab 50 + 1000 mg 1046 2.2 87.22/1744.35 93.33 0.83 1.2:1 64.6 (10.7)
Tab 50 mg 338 0.7 70.6 93.33 1.8:1 69.0 (11.9)
Vildagliptin + metformin, tab 50 + 850 mg 295 0.6 83.39/1417.69 93 1.4:1 64.7 (13.1)
Vildagliptin + metformin, tab 50 + 500 mg 129 0.3 76.96/769.64 93.30 1.3:1 66.8 (11.3)

Linagliptin
Tab 5 mg 848 1.8 4.8 5 0.92 1.1:1 70.2 (12.4)
Linagliptin + metformin, tab 2,5 + 1000 mg 352 0.7 4.23/1692.87 5 1.1:1 64.0 (11.6)
Linagliptin + metformin, tab 2,5 + 850 mg 151 0.3 4.31/1464.83 5 1.1:1 63.4 (12.1)
Linagliptin + metformin, tab 2,5 + 500 mg 67 0.1 3.83/765.34 5 1.0:1 70.2 (10.1)

Saxagliptin
Saxagliptin + metformin, tab 2,5 + 1000 mga 254 0.5 3.93/1573.93 5 0.88 1.4:1 63.3 (11.7)
Saxagliptin + metformin, tab 5 + 1000 mga 175 0.4 5.15/1029.84 5 1.9:1 63.9 (11.7)
Tab 5 mg 125 0.3 4.3 5 2.0:1 67.6 (12.1)
Tab 2.5 mg 30 0.1 2.1 2 0.7:1 70.5 (11.3)

Acarbose tab 50 mg 4 0.0 62.5 50 0.21 1.0:1 59.9 (9.3)

Dapagliflozin tab 10 mg 24 0.1 5.4 3.1 0.54 2.0:1 58.9 (11.6)
DDD: average ratio between the prescribed daily dose and the defined daily dose; SD: standard deviation.
a
Extended release presentation.

3.4. Economic analysis metformin is $1,540,931 USD. The annual dispensation of the
850 mg formulation of metformin (only included in the POS,
It was estimated that on average 8.2 DID of metformin were used by 33,898 patients) costs $236,145 USD, whereas formu-
consumed. The estimated annual cost of the dispensation of lations outside of the POS cost $1,304,786 USD. The cost per
 p r i m a r y c a r e d i a b e t e s 1 2 ( 2 0 1 8 ) 184–191
tablet varies between $0.01 USD COP and $1.05 USD depend-
ing on the formulation and the laboratory. The average cost
of a prescribed daily dose was $0.11 USD. The CID for the
metformin included in the POS was $0.19 USD and for all for-
mulations of metformin was $1.21 USD.
It was estimated that on average, 2.4 DID of glibenclamide
were consumed. The annual cost of the dispensation of gliben-
clamide to the affiliated population was estimated at $27,355
USD. The average cost of a prescribed daily dose was $0.01
USD, and the CID was $0.02 USD.
It was estimated that on average, 5.7 DID of insulins were
consumed. The estimated annual cost of human insulin was
$198,355 USD, whereas that of the analogs was $4,746,780 USD.
The average cost per unit of insulin was $0.002 USD and $0.02
USD for human insulin and the insulin analogs, respectively.
The CID for all insulins was $3.89 USD.
The total estimated cost of the dispensation of NPH insulin
was $161,776 USD for one year. The average cost of a prescribed
daily dose was $0.10 USD, and the CID was $0.13 USD. The total
cost of the dispensation of insulin glargine in one year was
calculated at $2,670,016 USD. The average cost of a prescribed
daily dose was $0.94 USD, and the CID was $2.10 USD.
4. Discussion
This study allowed us to define the prescription patterns of
antidiabetic medications in a population affiliated with the
Colombian Health System. This information can be used to
make decisions to improve the health care of DM patients.
The three main antidiabetic medications analyzed in this
study (metformin, glibenclamide, and insulins) are included in
the Colombian List of Essential Medicines and have been the
first choice for the management of DM in different phases of
the natural history of the disease. In our population, the rela-
tionship of the prescribed daily dose and the DDD was 0.75 and
0.85 for metformin and glibenclamide, respectively. This find-
ing showed that our patients were receiving lower doses than
recommended, although the doses were within the range sug-
gested by international consensus. This dosing may not only
affect adequate glycemic control but can also prevent patients
from obtaining the full benefits of the medication [9,11].
Especially for metformin, it is difficult to attain the DDD
of 2 g because the formulation is dispensed as 850 mg tablets.
This DDD could be achieved by three tablets a day, which could
reduce adherence to treatment due to increased adverse reac-
tions, which are primarily gastrointestinal in nature [12,13].
Moreover, the ratio of the prescribed daily dose and the DDD
of insulins in general was 1.30 ± 0.87, indicating great variabil-
ity in their prescription. This variability is probably related to
the fact that this drug must be adjusted in an individualized
manner [9].
The prescription rate for metformin to DM patients showed
an increase from 67.5% in 2005 to 81.3% in this study [5]. The
prescription frequency of this medication is highly variable
among the different studies conducted (from 48% in Japan
[14], 53% in the United States [15,16], 66% in Italy [17], 74% in
Taiwan [18], 84% in Canada [19], or even reaching 90% in the
United Kingdom primary care [20]). In the United States, met-
formin use also increased in recent years, but at a much lower
189
rate [21]. On the other hand, some countries have reported
a decreasing trend in metformin prescription, especially in
monotherapy, as shown in Austrian database analysis [22].
The prescription prevalence for sulfonylureas in Colombia
decreased significantly in the last 10 years, from 64.9% in 2005
[5] to 21.8% in this report. Studies published in other regions
(such as Canada, United States or Taiwan) also reported reduc-
tions, with prescription frequencies of 4.5%–30.8% [18,19,21].
This trend to lower sulfonylureas prescription might be related
to the emergence of newer antidiabetics with a better safety
profile (less hypoglycemia) [23] and this direction will probably
continue, especially in Colombia because the use of gliben-
clamide is no longer encouraged in recent guidelines [24].
The combinations most used in our population were
metformin–glibenclamide and metformin–insulin, which cor-
respond to accepted therapeutic indications and are similar to
those described by other authors [5,9,20,25].
The frequency of the use of metformin and insulin as
monotherapy can be considered adequate. In the case of DPP-4
inhibitors and glibenclamide, their high usage as monother-
apy does not seem justified because in the majority of cases
the use of medications that improve peripheral sensitivity to
insulin should be considered [11,26–28]. In Italy, thiazolidine-
diones had a prescription rate of 5.2% in 2010. However, in
this study, no patients with dispensation of these medications
were found, probably due to the alerts for cardiovascular risk
associated with their use [27,29].
General insulins use (33%) is higher than values close to
20% from Italy [17] but lower than the 41% reported in Austria
[22]. Compared with previous data from Colombia, insulins
presented an increase in prescriptions close to 10% [5], and
significant changes were observed in the type of insulin pre-
scribed. For instance, the prescription of human insulins was
reduced and the analogs grew, as was reported in other studies
[17,27]. In our country, this trend may be due to the inclusion of
these medications in the benefit plan, improving the accessi-
bility to these molecules without the need of a special medical
justification [30].
In 2011, global sales of insulins reached $16.7 billion USD.
This cost is rising due to the increasing trend in its prescription
and its increasing use in patients with DM. According to some
reports, insulin analogs are up to 126% more expensive than
human insulins, such as NPH [31]; in this study, the analogs
were 10.4 times more expensive. A study in Germany also
found that insulin therapy in general is linked with increased
annual medication costs [32]. Insulin analogs are promoted
with a better pharmacokinetic profile than human insulins;
however, they have failed to demonstrate a greater impact on
metabolic control among patients [31]. Other studies report
benefits in adherence, satisfaction, and quality of life, as well
as better control of glycated hemoglobin and fewer complica-
tions associated with severe hypoglycemia [33–35], although
investigations conducted in the Colombian population found
no significant differences in glycated hemoglobin, in the fre-
quency of hypoglycemia [36] or in the quality of life of patients
treated with conventional insulin compared to the analogs
[37].
Arterial hypertension and dyslipidemia were the most fre-
quent comorbidities (79.7 and 65.5%, respectively), which was
expected in this type of patient [38,39]. A considerable pre-
190 p r i m a r y c a r e d i a b e t e s 1 2 ( 2 0 1 8 ) 184–191
scription of analgesic and antiulcer drugs is also present,
which is similar to previous studies [5,27]. The prescription
of aspirin has undergone a substantial change (from 2.8% in
2005 [5] to more than 50% in the current study), although a
higher percentage would be expected considering the high
cardiovascular risk of patients with diabetes [39].
This research presented certain limitations for the inter-
pretation of some results because the information was
obtained from databases and no clinical records were con-
sulted to confirm the diagnosis or the therapeutic indication.
This study includes a specific population that receives medica-
tions included on a particular list. Therefore, the conclusions
are applicable to groups with similar insurance characteris-
tics and health care. Another important limitation is the lack
of data of therapy effectiveness, especially considering that
recent analysis showed that the proportion of patients achiev-
ing glycemic control did not improve over the last years in spite
of the increased use of insulin and newer costlier medications
[21,40].
Based on the prescription patterns found in this study, we
can affirm that generally rational prescription habits with the
use of drugs of high therapeutic value predominate. However,
the main drug (metformin) is used, on average, below the
internationally defined doses, many patients could probably
benefit from changes to drugs with lower risk of hypoglycemia
(particularly those using glibenclamide) and there is a large
number of patients with high-cost formulations, especially
insulin analogs. The patterns of use have changed notably over
time, and more changes would be expected as new molecules
are appearing on the market, making pharmacoepidemiolog-
ical studies an important source of data and suggesting the
necessity for new studies that also include cost-benefit anal-
ysis as well as complications and effectiveness variables.
The results of this study can be useful for system adminis-
trators and health services. Our findings should enable them
to make decisions about the way in which their doctors treat
this group of patients and to improve use patterns, thereby
reducing acute and long term complications.
Conflict of interest
The authors state that they have no conflict of interest.
Sources of funding
This study received funding from the Universidad Tecnológica
de Pereira, and Audifarma S.A.
Contributors
JEMA participated in the drafting, data collection, data anal-
ysis, description of results, discussion, critical revision of the
article, and evaluation of the final version of the manuscript.
DAMM, JASD and AGM participated in the drafting, data col-
lection, data analysis, description of results and discussion.
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at https://doi.org/10.1016/
j.pcd.2017.11.002.
references
[1] L. Guariguata, D. Whiting, I. Hambleton, et al., Global
estimates of diabetes prevalence for 2013 and projections for
2035, Diabetes Res. Clin. Pract. 103 (2) (2014) 137–149.
[2] J. Beagley, L. Guariguata, C. Weil, A.A. Motala, Global
estimates of undiagnosed diabetes in adults, Diabetes Res.
Clin. Pract. 103 (2) (2014) 150–160.
[3] M. Agudelo-Botero, C.A. Dávila-Cervantes, Carga de la
mortalidad por diabetes mellitus en América Latina
2000–2011: los casos de Argentina, Chile, Colombia y México,
Gac. Sanit. 29 (3) (2015) 172–177.
[4] P. Aschner, Epidemiología de la diabetes en Colombia, Av.
Diabetol. 26 (2) (2010) 95–100.
[5] J.E. Machado Alba, J.C. Moncada Escobar, G. Mesa Escobar,
Patrones de prescripción de antidiabéticos en un grupo de
pacientes colombianos, Rev. Panam. Salud Publica 22 (2)
(2007) 124–131.
[6] H. Vargas-Uricoechea, L.A. Casas-Figueroa, An epidemiologic
analysis of diabetes in Colombia, Ann. Glob. Health 81 (6)
(2015) 742–753.
[7] J.E. Machado-Alba, M.E. Machado-Duque, P.A.
Moreno-Gutierrez, Time to and factors associated with
insulin initiation in patients with type 2 diabetes mellitus,
Diabetes Res. Clin. Pract. 107 (3) (2015) 332–337.
[8] D.M. Nathan, The diabetes control and complications
trial/epidemiology of diabetes interventions and
complications study at 30 years: overview, Diabetes Care 37
(1) (2014) 9–16.
[9] American Diabetes Association, Approaches to glycemic
treatment, Diabetes Care 39 (Suppl. 1) (2016) S52–S59.
[10] N. Bansal, R. Dhaliwal, R.S. Weinstock, Management of
diabetes in the elderly, Med. Clin. North Am. 99 (2) (2015)
351–377.
[11] Asociación Latinoamericana de Diabetes, Guías ALAD sobre
Diagnóstico, control y tratamiento de la Diabetes Mellitus
tipo 2 con medicina basada en evidencia, Rev ALAD, 2013,
17–8.
[12] American Diabetes Association, 1. Strategies for improving
care, Diabetes Care 39 (Suppl. 1) (2016) S6–S12.
[13] L. Tran, A. Zielinski, A.H. Roach, et al., Pharmacologic
treatment of type 2 diabetes: oral medications, Ann.
Pharmacother. 49 (5) (2015) 540–556.
[14] M. Oishi, K. Yamazaki, F. Okuguchi, et al., Changes in oral
antidiabetic prescriptions and improved glycemic control
during the years 2002–2011 in Japan (JDDM32), J. Diabetes
Investig. 5 (5) (2014) 581–587.
[15] M.R. Sargen, O.J. Hoffstad, D.J. Wiebe, et al., Geographic
variation in pharmacotherapy decisions for US medicare
enrollees with diabetes, J. Diabetes Complications 26 (4)
(2012) 301–307.
[16] L.W. Turner, D. Nartey, R.S. Stafford, et al., Ambulatory
treatment of type 2 diabetes in the US, 1997–2012, Diabetes
Care 37 (4) (2014) 985–992.
[17] M. Baviera, L. Monesi, I. Marzona, et al., Trends in drug
prescriptions to diabetic patients from 2000 to 2008 in Italy’s
Lombardy Region: a large population-based study, Diabetes
Res. Clin. Pract. 93 (1) (2011) 123–130.
 p r i m a r y c a r e d i a b e t e s 1 2 ( 2 0 1 8 ) 184–191
[18] C.H. Liu, S.T. Chen, C.H. Chang, L.M. Chuang, M.S. Lai,
Prescription trends and the selection of initial oral
antidiabetic agents for patients with newly diagnosed type 2
diabetes: a nationwide study, Public Health 152 (2017) 20–27.
[19] A.S. Abdelmoneim, D.T. Eurich, J.-M. Gamble, S.H. Simpson,
Use patterns of antidiabetic regimens by patients with type
2 diabetes, Can. J. Diabetes 37 (6) (2013) 394–400.
[20] A. Maguire, B.D. Mitchell, J.C. Ruzafa, Antihyperglycaemic
treatment patterns, observed glycaemic control and
determinants of treatment change among patients with
type 2 diabetes in the United Kingdom primary care: a
retrospective cohort study, BMC Endocr. Disord. 14 (2014) 73.
[21] K.J. Lipska, X. Yao, J. Herrin, et al., Trends in Drug utilization,
glycemic control, and rates of severe hypoglycemia,
2006–2013, Diabetes Care (2006), dc160985.
[22] L. Lunger, A. Melmer, W. Oberaigner, M. Leo, M. Juchum, K.
Polzl, et al., Prescription of oral antidiabetic drugs in Tyrol —
data from the Tyrol diabetes registry 2012–2015, Wien. Klin.
Wochenschr. 129 (1) (2017) 46–51.
[23] B.M. Mishriky, D.M. Cummings, R.J. Tanenberg, The efficacy
and safety of DPP4 inhibitors compared to sulfonylureas as
add-on therapy to metformin in patients with Type 2
diabetes: a systematic review and meta-analysis, Diabetes
Res. Clin. Pract. 109 (2) (2015) 378–388.
[24] Ministerio de Salud y Protección Social, Guía de práctica
clínica para el diagnóstico, tratamiento y seguimiento de la
diabetes mellitus tipo 2 en la población mayor de 18 años,
Guía para profesionales de la salud, Colombia, 2015. [cited
October 2016]. Available: http://gpc.minsalud.gov.co/
gpc sites/Repositorio/Conv 637/GPC diabetes/GPC diabetes
tipo2 completa.aspx.
[25] A.S. Geier, I. Wellmann, J. Wellmann, et al., Patterns and
determinants of new first-line antihyperglycaemic drug use
in patients with type 2 diabetes mellitus, Diabetes Res. Clin.
Pract. 106 (1) (2014) 73–80.
[26] B. Banu, M. Shahi, K. Begum, et al., Prescribing behavior of
diabetes treating physicians in selected health care facilities
of the Diabetic Association of Bangladesh, Indian J. Public
Health 58 (3) (2014) 180–185.
[27] M. Baviera, L. Cortesi, M. Tettamanti, et al., Changes in
prescribing patterns and clinical outcomes in elderly
diabetic patients in 2000 and 2010: analysis of a large Italian
population-based study, Eur. J. Clin. Pharmacol. 70 (8) (2014)
965–974.
[28] D. Jayawardene, G.M. Ward, D.N. O’Neal, et al., New
treatments for type 2 diabetes: cardiovascular protection
beyond glucose lowering? Heart Lung Circ. 23 (11) (2014)
997–1008.
191
[29] Instituto Nacional de Vigilancia de Medicamentos y
Alimentos, INVIMA suspende en Colombia la
comercialización y uso del medicamento rosiglitazona,
Alerta sanitaria INVIMA 006-10, 2010. [cited October 2016]
Available: https://www.invima.gov.co/images/pdf/Prensa/
alerta-sanitaria/Alerta 06 29-sept Suspension Rosiglitazona.pdf.
[30] J.E. Machado-Alba, D. Torres, A. Portilla, A. Felipe Ruiz,
Results of the inclusion of new medications in the obligatory
health system plan in Colombia, 2012–2013, Value Health
Reg. Issues 8 (2015) 28–35.
[31] M.B. Davidson, Insulin analogs—is there a compelling case
to use them? No!, Diabetes Care 37 (6) (2014) 1771–1774.
[32] L. Jacob, C. von Vultee, K. Kostev, Prescription patterns and
the cost of antihyperglycemic drugs in patients with type 2
diabetes mellitus in Germany, J. Diabetes Sci. Technol. 11 (1)
(2017) 123–127.
[33] G. Grunberger, Insulin analogs—are they worth it? Yes!,
Diabetes Care 37 (6) (2014) 1767–1770.
[34] P. Rys, P. Wojciechowski, A. Rogoz-Sitek, et al., Systematic
review and meta-analysis of randomized clinical trials
comparing efficacy and safety outcomes of insulin glargine
with NPH insulin, premixed insulin preparations or with
insulin detemir in type 2 diabetes mellitus, Acta Diabetol. 52
(4) (2015) 649–662.
[35] F. Wang, J.M. Carabino, C.M. Vergara, Insulin glargine: a
systematic review of a long-acting insulin analogue, Clin
Ther. 25 (6) (2003) 1541–1577.
[36] A. Rivera, T. Álvarez, F. Ochoa, Comportamiento de
hemoglobina glicosilada y frecuencia de hipoglicemias en
diabéticos tipo 2 tratados con insulina glargina o NPH, Rev.
Méd. Risaralda 21 (2) (2015) 3–10.
[37] J.E. Machado-Alba, D.A. Medina-Morales, L.F.
Echeverri-Cataño, Evaluation of the quality of life of patients
with diabetes mellitus treated with conventional or
analogue insulins, Diabetes Res. Clin. Pract. 116 (2016)
237–243.
[38] P. Aschner, Diabetes trends in Latin America, Diabetes
Metab. Res. Rev. 18 (Suppl. 3) (2002) S27–S31.
[39] American Diabetes Association, 8. Cardiovascular disease
and risk management, Diabetes Care 39 (Suppl. 1) (2016)
S60–S71.
[40] V. Higgins, J. Piercy, A. Roughley, et al., Trends in medication
use in patients with type 2 diabetes mellitus: a long-term
view of real-world treatment between 2000 and 2015,
Diabetes Metab. Syndr. Obes. 9 (2016) 371–380.