Possible pharmacological interactions in hypertensive and / or diabetic elderly in family health units at Blumenau ( SC )

The aim of this study was to examine the possible adverse drug-drug interactions in the elderly population (≥60 years) diagnosed either with diabetes, arterial hypertension or both, at a Family Health Unit (FHU) clinic in Blumenau, state of Santa Catarina, Brazil. For this purpose 318 subjects were interviewed using a pretested form with social and demographic aspects regarding their disease and its complications. All drugs used by this population were grouped, and the drug-drug interactions were detected by using the “Drug Interaction Checker” software, and classified for frequency and severity. The average age of patients was 70.6 years, with a higher number of female patients (216; 67.9%). Most subjects were being treated for both diseases (149; 46.86%). Out of a total of 1,541 medications prescribed, the most prevalent were: hydrochlorothiazide (131; 8.4%) and enalapril maleate (130; 8.4%). A total of 295 possible drug-drug interactions were detected in 152 patients (1.9 interactions per subject), 275 (93.2%) moderate and 20 (6.8%) severe or highly severe. The possible interaction prevalence was 0.93 (0.55 to 1.40 depending on health unit). The most frequent possible interaction was that between acetylsalicylic acid and enalapril maleate (37; 12.5%). Patients had an average consumption of 6.6 drugs and 9.8% of subjects reported physical discomfort when using medicines. Elderly patients use many medications that could cause adverse reactions and possible drug-drug interactions, where this issue warrants closer attention of prescribers and health providers.

Uniterms: Drug-drug interaction.Systemic arterial hypertension/use of medications.Diabetes Mellitus/ use of medications.Elderly/use of medications.Family health program.

INTRODUCTION
Weight gain in the Western population as a whole, together with inadequate feeding habits, physical inactivity, and the use of tobacco and alcohol, have contributed to an increase in the prevalence of health problems such as arterial hypertension and diabetes (Ogden et al.., 2006).
In the presence of this unbalanced way of life, physicians are facing a great challenge: providing adequate therapeutic support to patients with interrelated and even synergistic diseases.These resultant drug combinations can decrease the overall treatment efficacy and/or favor adverse reactions of varying degrees of severity.The strategy to treat coexisting pathologies results in polypharmacy (use of five or more drugs), a situation that could be exacerbated by the consumption of over-the-counter (OTC) remedies (Junius-Walker et al., 2007).
With the increased life expectancy in Brazil, the at-risk population for development of diverse age-related diseases has grown.As a result, clinicians are faced with a growing number of patients who are evidently fragile to deal with invasive treatments and interventions.Thus, polypharmacy appears to be a problem in this sub-population.In a review, Rozenfeld (2003) pointed out that Brazilian epidemiological studies show that medicine use increases with age, and around 80% of aged people consume at least one medicine regularly while 30% of elderly consume over-the-counter (OTC) medicines.Medical prescriptions in this group commonly evidence inadequate doses and indications, unexpected pharmacological interactions and unsuitable medication associations, besides the use of therapeutically ineffective medications.
The treatment of patients with arterial hypertension and/or diabetes is no exception regarding polypharmacy.When patients are diagnosed with arterial hypertension and/or diabetes, many different medications could be appropriate for their treatment, including drugs for dyslipidemia, hypertension, antiplatelet therapy and glycemic control (Triplitt, 2006).
The identification of drug-drug interactions in patients with hypertension and/or diabetes could be very important, because these diseases are highly prevalent in Brazil.The treatment of these diseases includes highly complex therapeutics, which are undergoing rapid development and employ substances that are relatively new in the market, whose interactions are not yet fully clarified, or documented.A significant proportion of these new drugs are available on the Brazilian public health system (Brasil, 2009).
The extent of suffering and deaths caused directly by pharmacological interactions is not clear, but cur-rent evidence shows that this problem deserves serious attention on the part of both the medical community and the population who self-medicate and disregard the risks involved (Lima, 2004).Currently, an increasing number of people have access to a host of medications (thanks to large number of medication providers, such as physicians, pharmacists, friends and relatives), and their utilization can become abusive, particularly with regard to the risk for undesirable pharmacological interactions among chronically diseased patients (Lima, 2004;Rozenfeld, 2003).
Results from different studies on adverse drug reactions (ADRs) have been pooled in some metaanalyses, showing that the proportion of drug-related hospitalization is between 2.4 and 6.2% where many such cases were preventable (Cresweel et al., 2007).Another prospective observational study in England found that adverse drug reactions cost £466 million annually, resulting in the deaths of around 5700 patients (Pirmohamed et al., 2004).
The aim of the present study was to verify the possible pharmacological interactions in elderly arterial hypertension and/or diabetes patients attended at Family Health Strategy units of Blumenau, in the state of Santa Catarina, Brazil.

MATERIAL AND METHODS
This is a cross-sectional epidemiologic study on a sample population composed of elderly people with arterial hypertension and/or diabetes, living and registered in the area covered by the Family Health Strategy (FHS), in the city of Blumenau, state of Santa Catarina (SC), Brazil.
Elderly persons were considered individuals, aged 60 or older, as proposed by WHO for developing countries (Brasil, 2006).
Individuals with at least six months of clinical follow-up, and living in areas covered by FHS units in place for at least 1 year before the field work started in 2006 April, were included in the study.
Two-stage random sampling took place.In the first stage, a simple random sampling of 10 Family Health units was carried out (of the 34 existing units that had been in operation for more than 1 year).In the second sampling stage, those people living in the area covered by the units and registered as arterial hypertensive and/or diabetic were identified.Subsequently, stratified random sampling was done, distributing the sample size proportionally for people with arterial hypertension and/or diabetes registered in every unit.Thus, the people aged 60 years or more were identified and included in this study.
The study participants were visited at home and a pretested questionnaire was applied by trained and qualified interviewers.Data on the following variables were obtained: sex, age (average value and age ranges), marital status (if married, single, widow, or separated), race/ color (as reported by the interviewee), religion practiced, educational level (collected as completed years of study and categorized as 1 st to 4 th elementary school, 5 th -8 th middle school, and complete high school or graduated), current consumption of tobacco and alcohol (type and quantity), time of diagnosis (in months), time of treatment (in months), other associated diseases, name and amount of consumed medications, use of natural homemade remedies, presence of any physical discomfort when using medication.
In order to estimate the presence of drug-drug interactions, the medications consumed by every person in the last 7 days were identified.These medications were entered into an on-line software program (Drug Interaction Checker), which performed a bibliographical search of three (3) databases: the U.S. Food and Drug Administration (FDA), American Hospital Formulary Service (AHFS Drug Information) and First DataBank, generating a report containing the following information: 1) Presence of interaction: whether any type of interaction is occurring or not; 2) Severity: any interaction being classified as 'very severe' (drug combination is contraindicated and should not be dispensed or administered to patient); 'severe' action is required to decrease the interaction risk); and 'moderate' (interaction risk should be evaluated and action measures are also necessary); 3)Symptoms and clinical signs and; 4) Mechanism of action.
Consumed medications were organized according to the Brazilian Common Denomination, which assigns the name of active principles utilized by pharmaceutical industry through the Ordinance # 1179, dated June 17 th , 1996.
The data were descriptively presented in tables by frequencies with a confidence interval of 95%.For the continuous variables, the central trend (mean and median) and dispersion (standard deviation) measures were calculated.The prevalence (number of events divided by persons at risk) of possible interactions was also calculated.
This present study was developed with data collected by the Project "Adherence to pharmacological treatment of patients with hypertension and/or diabetes mellitus in the family health units in Blumenau (SC)", previously approved by the Institutional Review Board/ Independent Ethics Committee of the Regional University of Blumenau (040/2004).

RESULTS
Three hundred and eighteen individuals aged 60 years or older with arterial hypertension and/or diabetes, were interviewed.
These people were predominantly female (216; 67.9%), had a mean age of 70.6 years (standard deviation of 6.9 years) and median age of 69.9 years.The majority of these people had studied only up to 1 st -4 th years of elementary school (210; 66.2%) or were illiterate (75; 23.7%) (Table1).
Table 2 shows information on diseases and medications.
The average treatment time of patients with arterial hypertension was 162.3 months (SD = 144.7)and median was 120 months.The average time for treatment of diabetes was 115.3 months (SD = 96.6)and the median was 84 months.
The presence of 295 possible drug-drug interactions in 152 elderly people (47.8%; n= 318) was detected.Out of these possible interactions, 275 were of moderate severity (93.2%), 19 were severe (6.5%) and 1 was very severe (0.3%).The prevalence of possible interactions was 0.93, (0.55 to 1.40 depending on the family health unit).
Twenty possible interactions were detected with a higher grade of severity.One of these was considered very severe and occurred with nortriptyline hydrochloride and selegiline.This drug-drug interaction can lead to a severe reaction, including the symptoms: hyperpyrexia, convulsions, excitability, muscle stiffness, fluctuations in blood pressure, epilepsy, coma, and death.
The clinical manifestations and mechanism of action of more severe pharmacological interactions are shown in Table VI.It is noteworthy that in one individual, 11 possible drug-drug interactions were detected, 3 being severe and 8 moderate.The severe interactions comprised: amiodarone hydrochloride and digoxin; amiodarone hydrochloride and sodium warfarin; sulfasalazine and sodium warfarin.
Potential drug-drug interactions were detected in 152 out of the 318 elderly interviewed (47.7%).This frequency is lower than the figure reported in a Mexican study, where almost 80% of patients presented potential pharmacologic interactions.This difference could be associated to the Mexican study's inclusion criteria, namely, patients older than 50 years, with non-malignant algic syndrome.It is noteworthy that the prevalence of hyperten-  sive and diabetic patients was higher in the present study (67.3% and 29.5%, respectively), as were the predominant interactions, generally associated to medications used to treat these pathologies (Dubova et al., 2007).
In addition, another European multicenter study conducted in 5 countries involving individuals aged 65 years or older in use of 4 medications or more, detected some type of possible interaction in 46% of the sample (Bjorkman et al., 2002).
With respect to tobacco use, the vast majority, 90.9% of those interviewed, stated they were non-smokers.Studies have shown that the percentage of smokers declines proportionally with age (Paniz et al.., 2008;Sawatzky et al.., 2007;Dubova et al., 2007).In a comparative study of

Sodium Warfarin Sulfasalazine
Bleeding risk; Decreased Prothrombin activity.

Sodium Warfarin Sodium Levothyroxine
Hypothyroidism could increase the oral anticoagulant requirement.
The administration of thyroid hormones or hyperthyroidism could decrease the oral anticoagulant requirement.

Diazepam Fluoxetine Hydrochloride
Increased BZD levels; Clinical effect increase of BZD.
Certain SSRIs (Selective Serotonin Reuptake Inhibitors) could decrease the Phase I metabolism of BZD through competitive inhibition of CYP-450.
smoking prevalence, a higher decline in total percentage of smoking elderly individuals was evident from 1989 until 2003 where in 2003 a percentage of 14% of elderly smokers was found in the sample (Monteiro et al., 2007).
In this study, 60 persons (18.9%) reported physical discomfort with the use of some medication, with stomach pain (12%) and nausea (9.3%) being the predominant complaints.The inadequate use of medications in elderly has been high in many studies, with some symptoms being connected with specific drugs.In the study of Gurwitz et al .. (2003) there was association of dry mouth with the use of antidepressive medications and nauseas with digoxin and NSAIDs.According to Dubova et al.. (2007), cardiovascular disease allied to polypharmacy (>5 drugs) and advanced age (>60 years), are risk factors for drug-drug interactions.
The most consumed medication was hydrochlorothiazide (8.4%) followed by enalapril maleate (8.4%), lovastatin (6.5%), metformin (6.3%), acetylsalicylic acid (5.6%), glibenclamide (5.6%) and atenolol (3.8%).Given this, with the exception of metformin and omeprazole, the 10 most-prescribed medications were associated with 58% of cases of possible interactions.The predominance of drugs affecting the cardiovascular system, pancreas and endocrine system and drugs utilized in homeostasis and thrombosis was also evidenced in a Brazilian study with elderly individuals, irrespective of the disease (Marin et al., 2008).
The most frequent potential drug-drug interactions include acetylsalicylic acid with enalapril maleate or glibenclamide, but most patients use acetylsalicylic acid in low doses (80 to 200mg) to prevent major cardiovascular events, thereby lowering the risk of interactions.Out of 295 possible drug-drug interactions, 20 (6.8%) were considered severe or very severe and included the most frequently prescribed medications (metformin, glibenclamide and atenolol).The use of norfloxacin it notable because, when associated to metformin (2 occurrences) or glibenclamide (2 occurrences) norfloxacin can cause deep hypoglycemia by an as yet unknown mechanism.Norfloxacin should also be prescribed with caution in elderly when associated to steroids because of the risk of tendon ruptures.The use of trimethoprim-sulfamethoxazole (in patients utilizing metformin) or cephalexin (in those using glibenclamide) could represent alternatives.
Also, it is important to consider the patient's renal function and manage this risk.The use of atenolol combined with clonidine (1 occurrence) could lead to severe hypertension by blockage of B2 receptors, annulling the balance with vasoconstriction mediated by alpha receptors.An alternative to this prescription could be the combination of atenolol with enalapril maleate (Mion et al., 2002).
Digoxin was the fourth medication in the list of main causers of pharmacological interactions, and was related to 13.8% of pharmacological interactions, despite being present in only 1.6% of prescriptions.Due to its narrow therapeutic window (0.8 to 2.0 ng/mL) and reduced renal function in elderly patients, there is a higher chance of toxicity.Its association with amiodarone represents 1% of interactions and is considered severe, causing effects of digitalis intoxication.It is a drug considered inappropriate, but of low risk by the Beers criterion for elderly, at doses above 0.125mg/dose, even after taking into account the presence of other drugs for cardiac failure (inhibitors of angiotensin conversion enzyme, betablockers and diuretics) and for arrhythmias (amiodarone) (Packer et al., 1993;Fick et al., 2003;Fialová et al., 2005;Chutka et al.., 2004).
Many studies are evaluating the use of electronic means to help healthcare professionals in the identification of possible pharmacological interactions in their prescriptions.The possibility of reducing pharmacological interactions using such devices is a current focus of study.However, many difficulties have been pointed out in the use of such systems, including the main issue of the lack of specificity in results, with too many interactions being indicated, few of which are actually relevant.Other problems include the lack of emphasis in warnings for the more severe interactions, leading to many people ignoring or underestimating the warnings issued by these computer programs (Tamblin et al., 2003;Ahearn, Stephen, 2003).

Limitations
This study presents some limitations.There are several definitions of elderly and most of these can be criticized because they are based on age alone.Although the WHO recommends 60 years and older for developing countries, many authors around the world have adopted 55 or 65 years.It is important to bear in mind that being elderly does not only refer to age alone, but to a special phase in life, all of which are related to particular conditions (social, clinical, emotional) that must be taken into account when comparisons are made There are different software programs supported by a variety of databases which identify possible drugdrug interactions.This array of options could explain the frequency of discrepancies found among studies.In this study, an open web database was used to facilitate access and achieve better comparisons.
It is worth mentioning that cross-sectional studies such as the present investigation, can only detect preva-lent cases of interactions.In severe cases, it is likely that patients with symptoms complain to their doctors or ask to change medicines.Therefore, some cases can only be detected in longitudinal investigations such as cohort studies.

CONCLUSIONS
Access to medications was broadened following implementation of the Family Health Strategy, particularly to drugs requested for hypertension and diabetes (Paniz et al., 2008).However, the high proportion of possible drug-drug interactions and adverse reactions suggest that measures should be taken in order to improve the quality of care, by means of continued education for prescribers, as well as by improvement of dispensation conditions in these units (usually done by nursing staff).The frequent staff turnover of family health teams, compromising sustained continuity of care, associated to treatment that is sometimes given by different medical specialists, could also contribute to the occurrence of possible interactions.Further investigations should be undertaken to evaluate the consistency of the results obtained and to study their possible determinants.

TABLE I -
Socio-demographic variables pharmacological interactions in hypertensive and/or diabetic elderly in family health units at Blumenau (SC) 799 Possible

TABLE II -
Treatment

TABLE IV -
Most frequent medications in interactions (n=590)

TABLE V -
Most frequent potential drug-drug interactions and their clinical effects

TABLE VI -
Clinical manifestations and mechanism of action of most severe potential drug-drug interactions