SEARCH

SEARCH BY CITATION

Keywords:

  • database;
  • drug-interactions;
  • elderly

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Acknowledgements
  10. Financial disclosure
  11. Sponsor role
  12. Conflict of interest
  13. Author contribution
  14. References

Purpose:  To estimate the prevalence of potentially severe drug–drug interactions (DDIs) and their relationship with age, sex and number of prescribed drugs.

Methods:  We analysed all prescriptions dispensed from 1 January 2003 to 31 December 2003 to individuals aged 65 or more registered under the Local Health Authority of Lecco, a northern Italian province with a population of almost 330 000 persons. Elderly who received at least two co-administered prescriptions were selected to assess the presence of DDIs.

Results:  The prevalence of potentially severe DDIs was 16%, and rose with increasing patient’s age and number of drugs prescribed. At multivariate analysis, the adjusted odds ratios rose from 1·07 (95% CI 1·03–1·11) in patients aged 70–74 to 1·52 (95% CI 1·46–1·60) in those aged 85 or older. Elderly taking more than five drugs on a chronic basis had a statistically significant higher risk of sever DDIs than those receiving less than 3 or 3–5 such drugs.

Conclusions:  The elderly constitutes a population at high risk of DDIs. As physicians still have some difficulty in managing this problem, it is essential to highlight for them, which factors raise the risk of DDIs.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Acknowledgements
  10. Financial disclosure
  11. Sponsor role
  12. Conflict of interest
  13. Author contribution
  14. References

Many studies have documented the rise in prescription of medicines for elderly patients and the related hazards, including difficulties in adherence to therapies, side-effects and drug–drug interactions (DDIs) (1–8). In Italy, nearly 20% of the population is aged 65 years or more, and 50–60% of overall drug consumption is attributable to them (9).

Drug-related problems and adverse drug reactions (ADRs) are important causes of morbidity and mortality; they are responsible for up to 6–7% of hospital admissions, and represent a significant burden in terms of healthcare costs (3–5, 8, 10). The risk of ADR rises with age and the number of drugs taken. The risk of a serious ADR is estimated at 19% in adults aged 55–64 years, 30% at 75–84 and more than 40% in patients aged 85 or older (8, 10). Although many ADR are unpredictable, others can be foreseen and prevented, such as those as a result of DDIs (11–13). Prevention is possible by focusing on previous reports, experimental or clinical studies, pharmacological principles or using computerized drug-interaction alert systems that screen prescriptions and flag potential interactions, and hence alert physicians about dangerous drug combinations (14–16).

Elderly patients are the population at highest risk of potential DDIs because of their higher exposure to polypharmacy, age-related physiological changes in pharmacokinetic and pharmacodynamic characteristics, impairment in many organ functions (particularly kidney and liver), and, not least, the number of physicians they consult (17, 18). In community-dwelling elderly, the prevalence of potential DDIs ranges from 4 to 46%, but the figures vary greatly in relation to the setting, population and system of DDI evaluation (7, 8, 12, 19–23). A recent study found a strong relationship between number of drugs and potentially severe DDIs (24).

In spite of this evidence, however, little information is available about the prevalence of severe DDIs in elderly people, because many studies have focused on preclinical settings, healthy volunteers, case reports, middle-aged or highly selected patients. The aim of this study was to estimate the prevalence of potentially severe DDIs in the elderly population registered under the Local Health Authority (LHA) of Lecco (Northern Italy) during the year 2003, and their relationship with age, gender and number of chronically prescribed drugs.

Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Acknowledgements
  10. Financial disclosure
  11. Sponsor role
  12. Conflict of interest
  13. Author contribution
  14. References

Study design

This observational, cross-sectional study is part of an ongoing pharmaco-epidemiological project, in collaboration with the Health Department of the Lombardy Region (a Northern Italian Region with more than nine million inhabitants) designed to estimate drug consumption and clinical outcomes of all community dwelling elderly residents in Lombardy during the year 2003. This report focused on data from the LHA of Lecco, one of the northern provinces of Lombardy with a population of almost 330 000, 275 general practitioners (GPs) and 37 family paediatricians (FPs).

All 58 800 individuals aged ≥65 years registered under the Lecco LHA from 1 January 2003 to 31 December 2003 were considered eligible for inclusion. All drugs prescribed for each elderly patient in the study period were considered. Elderly people who received at least two prescriptions (each of one generic entity) were selected to evaluate the presence of potential severe DDIs.

To assess the frequency and distribution of potentially severe DDIs, all drug prescriptions dispensed from 1 January 2003 to 31 December 2003 were considered. Every prescribed drug pair was evaluated for the potential risk of severe DDI using an Italian computerized interaction database (25).

Data source

The Lombardy Region Health Department collects prescription data monthly from all LHA of the region. These data are grouped in a regional database that can be linked to other administrative database (patients’ personal data, physicians’ data) using a specific identification code. These prescriptions refer to drugs prescribed by GPs and FPs dispensed by the Italian National Health Service, according to the Italian National Drug Formulary criteria. The regional database does not include information on over-the-counter drugs, herbal medicines and drug used in hospitals or nursing homes. The regional drug prescription database includes a full account of the product dispensed, the date of purchase, the personal identification code of each patient who received the prescription and of the prescribing physician. All prescriptions were classified according to the Anatomical Therapeutic Chemical (ATC) classification system, as recommended by the WHO (26).

Demographic data on patients (sex, date of birth, place of residence) and physicians (sex, age, place of residence and year of graduation) are available in regularly updated ad hoc regional databases, which can be linked through patient or physician identification keys.

Drug interaction database

Potential severe DDIs were analysed by a computerized system, using the Italian interaction database, an electronic version of an Italian textbook for clinical evaluation of DDIs (27). In this system, all drug interactions are classified in terms of clinical relevance, as severe, moderate or minor, taking into account ‘potential’ clinical outcomes, type, quality and relevance of supporting clinical and pharmacological documentation. Each potential DDI is classified by clinical relevance as: severe (drug combination should usually avoided or may potentially lead to serious clinical consequences, such as severe adverse effects or no clinical effects, close monitoring is required), moderate (drugs can be combined, the precipitant drug may modify the effect of the object drug, but the resulting effect can be controlled by individual dose adjustment and/or by controlling the plasma concentration of the drug) and minor (drug combination probably has no clinical relevance or has not been completely assessed). For each potential DDI, besides the classification of clinical relevance, the system provides the physician with information on the mechanism responsible for the interaction (if available), the clinical or pharmacological effect of the interaction, advice on measures to control or manage the risk of interaction, and the type and relevance of documentation available (preclinical studies, case reports, clinical studies, epidemiological studies). This information was collected from the ‘Summary of Product Characteristic’ of each drug, in a similar way to that used in other well-known DDI databases (e.g. Micromedex DRUG-REAX®) and by a review of available clinico-pharmacological literature. All DDIs refer to the chemical active substances (International Non-proprietary Name) and through them the system is linked to the ATC classification code and to the Italian National Drug Formulary.

The occurrence of a potential DDI was defined as the presence of at least a day of co-prescription of two drugs (object and precipitant drugs) within a period of at least 2 weeks, based on prescription delivery dates. Delivery dates indicate the dates on which the elderly received their prescriptions. Only the interactions identified as severe were considered in these analyses.

We also evaluated the role of exposure to chronic drug therapies, classified as the consecutive prescription of at least four packets of a drug (identified by the active substance) in the year of the study.

Before beginning the study, we tested the consistency of our drug interaction database in comparison with the widely used and well-known Micromedex DRUG-REAX® System (28).

Data analysis

Data were retrospectively analysed considering the period 1 January 2003 to 31 December 2003. We analysed periods of overlapping prescriptions for all individuals, and identified all combinations of potentially interacting drugs. DDIs were defined as two prescriptions that coincided. Searches were done both within prescriptions and across successive prescriptions for each patient over at least 2 weeks.

Potentially severe drug interactions were analysed in terms of number of patients involved and number of interactions. Data were stratified for sex, age and number of prescriptions. Logistic regression analysis was used to study the associations between number of chronic drugs and potentially severe DDIs, with adjustment for age and sex. The results are shown as odds ratios (OR) with 95% confidence intervals (CI). jmp software (SAS, Cary, NC, USA) version 7.0.1 for Mac OSX was used for the analyses.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Acknowledgements
  10. Financial disclosure
  11. Sponsor role
  12. Conflict of interest
  13. Author contribution
  14. References

General characteristics

Among the 326 130 residents under the Lecco LHA at 31 December 2003, the 58 800 (18·0%) aged 65 years or older were included in the study. Of these, 50 324 (85·6%) received at least one drug prescription during the year of study (84·5% of men and 86·3% of women). The characteristics of the elderly population included are shown in Table 1. The elderly patients received 974 002 drug prescriptions (50·5% of the overall 1 788 817 prescriptions), corresponding to a mean of more than four active substances per person. More than 70% were exposed to at least one chronic drug.

Table 1.   Characteristics of the elderly people and number of elderly with potentially severe drug–drug interactions (DDI) from the Local Health Authority of Lecco in 2003
Variablesn %
  1. aAt least four packs of a drug with the same active substance prescribed during 2003.

Elderly people (age ≥ 65 years)58 800100
 Women35 39360·2
 Age, mean (±SD)75·1 (7·3)
 Age group (years)
  65–6917 38829·6
  70–7414 94525·4
  75–7912 25520·9
  80–84808113·7
  ≥85613110·4
 At least one prescription50 32485·6
 At least one chronic prescriptiona43 23773·5
Number of elderly people exposed to five or more chronic prescriptions970316·5
Total prescriptions1 788 817100
Prescriptions to the elderly974 00250·5
Number of prescriptions/elderly person, mean (±SD)19·4 (16·9)
Total active substances692100
Active substances to the elderly59185·4
Number of active substances/elderly person, mean (±SD) 4·3 (3·8)
Number of chronic active substances/elderly person, mean (±SD)2·3 (2·4)
Elderly person58 800100
 With at least one potentially severe DDI942716·0
 With two potentially severe DDI28674·9
 With three potentially severe DDI10181·7
 With four or more potentially severe DDI6201·1
Overall number of potentially severe DDI13 932100
Number of DDIs/elderly person, mean (±SD)0·2 (0·7)

As expected, cardiovascular system drugs were the most frequently prescribed (68·5% of the elderly patients), followed by anti-infectives (44·6%), drugs for the musculo-skeletal system (42·8%), alimentary tract and metabolism (42·3%) and blood and blood forming organs (37·6%). Agents acting on the renin-angiotensin system (21% of all packs), calcium channel blockers (10%), cardiac therapy (9%), lipid-modifying agents (8%) and antithrombotic agents (7%) were the five most widely dispensed therapeutic groups.

Potentially severe DDIs

During the study, 9427 elderly people (16%) were exposed to drug combinations with the potential for 13 932 severe DDIs. Nearly 8% of the elderly population had two or more potentially severe DDIs and the mean number of interactions per patient was 0·2 (range 0–9) (Table 1).

The general characteristics of elderly people with at least one potential severe DDI and those without are shown in Table 2. Significant differences were found for age, numbers of prescriptions, packs and active substances, and rate of exposure to chronic drug therapies.

Table 2.   General characteristics of elderly patients with and without potentially severe drug–drug interactions (DDI)
VariablesElderly with DDIs n = 9427Elderly without DDIs n = 49 373P-value
  1. aChronic prescription: at least four pack of a drug with the same active substance prescribed during 2003.

Men, n (%)3729 (39·6)19 678 (39·9) 
Women, n (%)5698 (60·4)29 695 (60·1)0·59
Age, mean (±SD)77·5 (7·5)74·7 (7·1)<0·0001
Number of prescriptions, mean (±SD)35·0 (20·3)15·5 (13·6)<0·0001
Number of packs, mean (±SD)67·1 (39·9)25·1 (27·8)<0·0001
Number of active substances, mean (±SD)8·6 (4·1)3·5 (3·1)<0·0001
Elderly taking at least one chronic active substancea, n (%)9290 (98·6)33 947 (68·8)<0·0001
Number chronic active substancesa, mean (±SD)5·0 (2·7)1·8 (1·8)<0·0001
Elderly exposed to five or more chronic active substances, n (%)5022 (53·3)4681 (9·5)<0·0001

The 20 active substances most frequently involved in potentially severe DDIs are shown in Fig. 1. As expected, cardiovascular drugs were the most frequently involved (digoxin, enalapril, hydrochlorothiazide + amiloride, furosemide, allopurinol, etc.). Thirty-eight per cent of the 591 active substances prescribed to the elderly population were responsible for a potentially severe DDI.

image

Figure 1.  Active substances most frequently involved in potentially severe drug interactions.

Download figure to PowerPoint

The 13 932 potentially severe DDIs corresponded to 569 drug combinations. However, the first 20 (Fig. 2) were responsible for 42% of all potential DDIs. The combinations most frequently involved were digoxin and furosemide (6·4%), followed by nimesulide and acetylsalicylic acid (5·1%), allopurinol and enalapril (3·7%), and acetylsalicylic acid and diclofenac (3·2%).

image

Figure 2.  Drug–drug combinations (active substances) most frequently involved in potentially severe drug interactions.

Download figure to PowerPoint

Risk estimation

The prevalence of potentially severe DDI increased with increasing patient’s age (Fig. 3) and number of chronic drugs prescribed (Fig. 4). Both univariate and multivariate analyses show age and number of chronic drugs to be associated with an increasing risk of DDIs (Table 3). The adjusted OR increased from 1·07 (95% CI 1·03–1·11) in patients aged 70–74 years to 1·52 (95% CI 1·46–1·60) in those aged 85 or older. Furthermore, elderly taking more than five chronic drugs had a statistically significant higher risk of potentially severe DDIs (OR = 5·59; 95% CI 5·39–5·80) than those receiving less than 3 (reference category) or 3–5 chronic drugs (OR = 2·71; 95% CI 2·63–2·80). No statistically significant difference was observed in relation to sex (OR = 1·02; 95% CI 1·00–1·05).

image

Figure 3.  Distribution of the elderly resident under Lecco Local Health Authorithy with at least one potentially severe drug interaction in relation to sex and age.

Download figure to PowerPoint

image

Figure 4.  Distribution of the elderly resident under Lecco Local Health Authorithy with at least one potentially severe drug interaction in relation to sex and number of chronic drugs taken.

Download figure to PowerPoint

Table 3.   Odds ratios with 95% confidence intervals for potentially severe drug–drug interaction among elderly resident under the Lecco Local Health Authority in 2003
VariableUnivariate analysis OR (95% CI)Multivariate analysisa OR (95% CI)
  1. aAdjusted for sex, age and number of chronic drugs.

Age (years)
 65–6911
 70–741·20 (1·16–1·24)1·07 (1·03–1·11)
 75–791·40 (1·35–1·45)1·17 (1·12–1·21)
 80–841·54 (1·48–1·59)1·26 (1·21–1·32)
 85 or more1·74 (1·69–1·81)1·52 (1·46–1·60)
Sex
 Male11
 Female1·00 (0·98–1·03)1·02 (1·00–1·05)
Number of chronic drugs
 0–311
 3–52·77 (2·69–2·86)2·71 (2·63–2·80)
 6 or more5·71 (5·51–5·92)5·59 (5·39–5·80)

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Acknowledgements
  10. Financial disclosure
  11. Sponsor role
  12. Conflict of interest
  13. Author contribution
  14. References

In this study, 16% of elderly people registered under the Lecco Local Health Authority during the year 2003, were exposed to drug combinations, which put them at risk of potentially severe DDIs. Twenty-five per cent of the elderly aged 85 years or older was exposed to two or more drugs, and more than 60% of the elderly were exposed to six or more chronic therapies, and received combinations of drugs, which increased risk of potentially severe DDI. This risk increased with age and the number of chronic drug therapies co-prescribed.

A recently published study (24) reported similar results in people aged 75 years or older. The authors reported a strong relationship between the number of dispensed drugs and potential DDIs, especially severe ones. Unlike our results, they did not see any rise in risk with age, and elderly women had a lower probability of potentially serious DDIs than men. These differences might be related to differences in the prevalence of drug exposure of the oldest patients in the two populations. Our data considering only patients aged 75 years or older showed increasing risk of DDIs with age.

A relationship between age and potential DDIs was also reported in a Danish study, with the risk of DDIs rising from 24% in individuals aged 60–79 years to 36% in those over 80 (22). Another study of outpatients visiting a university hospital over a 12-month period also found similar trends: the rate of potentially significant DDIs increased with the patient’s age and prescription size (19) but was unaffected by gender.

The drugs we identified in potentially severe DDIs, are in agreement with previous reports (7, 8, 12, 20–23): cardiovascular, musculoskeletal, and blood-forming agents. Although acetylsalicylic acid occurred most frequently as a possible interactant (7·8% of the cases), in the majority of cases it was prescribed for the prophylaxis of cardiovascular and cerebrovascular events, at dosages of 100–300 mg/day. These are considered less likely to cause adverse reactions and drug interactions than the dosages normally used for the treatment of pain or musculo-skeletal diseases. As expected, the drugs most frequently responsible for potentially severe interactions were those widely used in primary care, including non-steroidal anti-inflammatory drugs, digitalis derivatives, anticoagulants, diuretics, antihypertensives and antimicrobial drugs.

We found that nearly 75% of elderly patients received at least one chronic drug and more than 16% were chronically treated with five or more drugs. Chronic therapies were associated with an increase in the risk of potential DDIs. Although no other studies specifically analysed the duration of therapies, we believe that the high prevalence of chronic therapies among elderly patients may be an important risk factor for DDIs. They may negatively influence the pharmacokinetics and pharmacodynamics of the drugs involved through induced changes in drug metabolism and excretion, and, not least, the body’s response to drugs (11, 29–32). Furthermore, chronic therapies often involve polypharmacy, another widely reported risk factor for DDIs (33–35).

The widespread use of medicines for elderly people, prescribed by different physicians, often unaware of other prescriptions or self-prescriptions by the patient, can significantly increase the risks of polypharmacy and hence of DDIs (17, 18). Polypharmacy is often related to chronic treatment with drugs for chronic diseases such as hypertension, diabetes, heart failure, arthritis, or respiratory and autoimmune disease (1, 2, 32–34). The risk of interactions may be underestimated because of the limitations of the majority of instruments available for assessing the probability of DDIs, which consider only single pairs of drugs and do not take into account interactions involving combinations of three, four or more (35, 36). Most of these instruments also do not take account of the dosages, duration of therapies and each patient’s risk factors.

Limitations

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Acknowledgements
  10. Financial disclosure
  11. Sponsor role
  12. Conflict of interest
  13. Author contribution
  14. References

The database we used to screen for DDIs refers to ‘potential DDIs’ and the classification of severity is based only on the potential seriousness of the clinical effect of the related event. Thus, we cannot know whether the hypothetical severity of a drug interaction resulted in any harm for the patients, or whether it had any negative influence on the therapeutic effects of one of the drugs involved. However, in many cases adverse effects can be prevented or avoided by dosage adjustment, administering drugs at different intervals, adequately monitoring the patient’s response or using drugs that do not interact. A study investigating the frequency, nature and side-effects of DDIs showed that among 538 elderly patients exposed to potential DDIs only 14·5% had a clinical and/or biological side-effect (23). A recent study comparing two methods for identification of DDIs [computerized screening (CS) and prospective bedside recording] showed that only 7% of DDIs detected by CS were recorded using the bedside approach (37). Although we are well aware of the clinical limitation and overstatement of DDIs using these kind of DDIs databases, this problem cannot be easily eliminated without use of a second ‘filter’ and clinical evaluation by the physician. It would be useful and helpful to set up specific algorithms for pooling information provided by the DDI database, to take account of dosage and duration of therapy, severity and frequency of the potential adverse events related to DDIs, and patient clinico-pharmacological characteristics and risk factors for drug-related problems (16).

Although our database has not been formally validated, we randomly selected a sample of 1000 patients from our cohort and screened them for DDIs using the DRUG-REAX® System (28) to evaluate the number, the type and the severity of potential DDIs. We did not find any significant difference and after the comparison of each DDI identified as ‘severe’ by our database with the level of severity attributed by DRUG-REAX® we found a concordance of more than 95% in the cases analysed.

Another weakness might be the inclusion in the analysis of all drugs prescribed chronically for the patients, regardless of when the treatment really started. Therefore, a patient considered at risk for a potential drug interaction might have been treated with the interacting drug for many years without any adverse effects.

Moreover, the source of prescribing data did not include information on over-the-counter drugs, dietary supplements, herbal medicines and drugs used in hospitals, thus possibly underestimating drug exposures and consequently the risk of potential DDIs. Furthermore, the lack of information on diseases or diagnoses of the patients included in the Lombardy Region Health Department Prescription database did not allow the study of consistency between the prescription of particular drugs and the presence of the pertinent diseases. However, physicians must prescribe these drugs only according to the criteria approved by the Italian National Drug Formulary in relation to specific diseases and diagnoses.

Finally, the use of administrative databases limits the number and type of variables (age, sex and number of prescribed drugs) that can be included in the analyses and the adjustment of data for important covariates such as comorbidities, functional and cognitive status.

Conclusions

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Acknowledgements
  10. Financial disclosure
  11. Sponsor role
  12. Conflict of interest
  13. Author contribution
  14. References

The elderly constitutes a population at high risk of potentially severe DDIs because of the high rate of drug prescription. This may be related to age-related clinical and pharmacological changes, as reported in other similar studies. As physicians still have difficulty managing this problem, it is essential to provide them with adequate information on risk factors for DDIs, and to help them in the evaluation of the clinical relevance of potential DDIs flagged by the computerized system. Although some related reports have already been made (15, 16), wider implementation in the clinical setting is still required. Physicians should be aware that elderly patients, especially the oldest group, those treated with polypharmacy for chronic diseases, those taking drugs with a steep dose–response curve and/or a narrow therapeutic index, and drugs metabolized by enzymes susceptible to induction or inhibition, have the highest risk of DDIs. In such cases, reviewing the patient’s drug record along with a clinical assessment of comorbidities, and presence of organ impairment, may help avoid adverse events as a result of DDIs. As recently reported (38), physicians should collect a detailed medication list for each patient, periodically review this drug regimen and adherence to it, adjust dosages when necessary, assess the appropriateness of prescribed drugs, educate patients or caregivers about the risk of polypharmacy and DDIs, and avoid prescribing inappropriate drugs (32).

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Acknowledgements
  10. Financial disclosure
  11. Sponsor role
  12. Conflict of interest
  13. Author contribution
  14. References

We thank Silvio Garattini and Carla Roncaglioni for reviewing the manuscript and providing useful suggestions, and Judith Baggott for editorial assistance.

Sponsor role

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Acknowledgements
  10. Financial disclosure
  11. Sponsor role
  12. Conflict of interest
  13. Author contribution
  14. References

The sponsors of the study had no role in the conception and design of the study; collection, management, analysis and interpretation of data; preparation and writing of the report or in the decision to submit the manuscript for publication.

Conflict of interest

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Acknowledgements
  10. Financial disclosure
  11. Sponsor role
  12. Conflict of interest
  13. Author contribution
  14. References

All the authors declare that no conflict of interest exists. All the authors state that they have a full control of data and that they agree to allow the journal to review their data if requested.

Author contribution

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Acknowledgements
  10. Financial disclosure
  11. Sponsor role
  12. Conflict of interest
  13. Author contribution
  14. References

Study concept and design: Nobili, Pasina, Tettamanti, Monesi, Marzona, Fortino, Bortolotti. Acquisition of data: Nobili, Pasina, Tettamanti, Cucchiani, Riva, Lucca, Monesi, Marzona, Fortino, Bortolotti, Merlino, Locatelli, Giuliani. Analysis and interpretation of data: Nobili, Pasina, Tettamanti, Cucchiani, Lucca, Monesi, Marzona, Riva, Lucca, Fortino, Bortolotti, Merlino, Locatelli, Giuliani. Drafting the manuscript: Nobili, Pasina. Critical review of the manuscript: Nobili, Pasina, Tettamanti, Lucca, Monesi, Marzona, Riva, Lucca, Fortino, Bortolotti, Merlino, Locatelli, Giuliani.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Acknowledgements
  10. Financial disclosure
  11. Sponsor role
  12. Conflict of interest
  13. Author contribution
  14. References