FormalPara Key Points

The prevalence of drug-related corrected QT (QTc) interval prolongation > 500 ms in geriatric medical inpatients may be higher than expected.

Proton pump inhibitors are largely prescribed in older inpatients with QTc interval prolongation.

Male sex is associated with long QT syndrome in older hospitalized patients.

1 Introduction

Long QT syndrome (LQTS) is a multifactorial form of heart rate (HR)-corrected QT (QTc) interval prolongation associated with an increased risk of torsade de pointes (TdP), a polymorphic ventricular tachycardia that can lead to ventricular fibrillation and sudden cardiac death [1]. A QTc interval > 500 ms is associated in the literature with a two- to threefold higher risk of TdP [1]. Recently, in a Norwegian study [2], patients with a QTc interval ≥ 500 ms had twice as high 30-day all-cause mortality as patients with a QTc interval < 500 ms, matched for age and sex and independent of comorbidity.

LQTS can be classified as either congenital or acquired. The acquired form of LQTS is far more prevalent than the congenital form and is predominantly related to the use of drugs blocking the human ether-a-go–go-related gene (hERG) potassium channel, thereby reducing the rapid component of the delayed rectifier potassium current IKr [3]. Other well-known risk factors of acquired LQTS include advanced age, hypokalemia, hypocalcemia, hypomagnesemia, bradycardia, organic heart disease, subarachnoid hemorrhage, starvation, and female sex [4]. However, two recent studies [5, 6] reported male sex as a risk factor of QTc interval prolongation in older inpatients. Moreover, specific autoantibody-mediated mechanisms occurring in patients with autoimmune diseases (anti-Ro/SSA antibodies-52kD subtype) [7] and systemic inflammation [8] have been identified as additional risk factors for TdP.

Aging is usually associated with polypharmacy, as older patients often need multiple medications, including QT-prolonging drugs, to treat comorbidities. A large German study reported that, at discharge from geriatric units, 58.7 and 22.1% of patients aged > 65 years received at least one QT-prolonging drug and two or more of such medications, respectively [9]. Moreover, a vast majority of geriatric patients are likely to cumulate other risk factors of TdP, such as heart failure, left ventricular hypertrophy, and electrolyte disturbances [10].

The prevalence of LQTS in patients admitted to different hospital settings ranges from 22 to 35% [4, 11,12,13,14,15]. Medical older inpatients are typically polymorbid and frequently receive QT-prolonging medications, so are at particular risk of TdP. A recent study utilizing data from a collaborative and independent prospective registry of the Italian Society of Internal Medicine (REPOSI), collecting data from 87 internal medicine and geriatric wards, found a prevalence of QT-prolonging drug use of 55.2% in geriatric inpatients on admission [16]. However, in this large retrospective analysis, Franchi et al. [16] did not collect electrocardiogram (ECG) data, electrolyte values, and other common risk factors of TdP with QT-prolonging drugs on admission or at discharge.

Despite the potentially high risk of drug-associated LQTS in geriatric inpatients, data in the literature are scarce. The primary objectives of this cross-sectional study were to estimate the prevalence of drug-related LQTS and the prevalence of use of QT-prolonging drugs in unselected older patients admitted to an internal medicine unit. Other aims of the study were to identify clinical and drug-associated risk factors of LQTS and to evaluate drug–drug interactions (DDIs) with a potential risk of QT prolongation (QT-DDIs) in this special population.

2 Materials and Methods

2.1 Patient Eligibility, Data Collection, and Measurement of the QTc Interval

One trained physician screened consecutive patients hospitalized in the Internal Medicine 2 Unit, University Hospital of Siena, from 13 April 2017 to 13 April 2019. Our institution is a public hospital with approximately 700 beds: our unit has 38 beds, with 86 hospitalizations per month recorded in 2017, 88 in 2018, and 92 in 2019. We excluded patients with atrial fibrillation, bigeminy, wide QRS complex (> 120 ms), or unidentifiable T waves on the baseline ECG as the QT interval cannot be reliably measured under such conditions. All patients aged > 65 years prescribed at least one drug on admission who signed the informed consent were enrolled in the study. At the time of enrollment, the physician collected information from patients on drugs prescribed by their general practitioner. Details of prescriptions during hospitalization were extracted from electronic medical records. Characteristics associated with LQTS in the medical literature [4] were also recorded in each case report form, including patient age, sex, body mass index (BMI), serum potassium, calcium, magnesium, and creatinine levels (all obtained the same day of the ECG) and diseases, including hypertension, hypothyroidism, type 2 diabetes mellitus, and cardiac diseases (acute coronary syndrome, dilated cardiomyopathy/heart failure, chronic ischemic cardiopathy, left ventricular hypertrophy). We defined hypokalemia as a potassium level < 3.5 mEq/L, hypocalcemia as calcium level < 8.0 mg/dL, hypomagnesemia as magnesium level < 1.50 mg/dL, and elevated serum creatinine as a creatinine level of > 1.10 mg/dL. Only levels obtained before the administration of intravenous magnesium sulfate and/or replacement therapy with potassium or calcium were considered appropriate for inclusion in the study. Within 24 h of patient admission, a trained physician performed a 12-lead ECG using an electrocardiograph with automated measurement of QT interval (Philips PageWriter TC50®). LQTS was defined as a QTc interval of > 450 ms in men and > 470 ms in women using Bazett’s formula (QTc = QT/√RR) [17]. Although the Bazett correction (QTcB) tends to produce overlong QTc interval values at faster heart rates [1], it was chosen because it is the most widely used in clinical practice, and QTc interval lengthening based on this correction formula has been associated with an increased risk of TdP and sudden death in the geriatric population [18].

As in most similar studies, we obtained the list of QT-prolonging medications from CredibleMeds® (formerly the Arizona Center for Education and Research on Therapeutics [AZCERT]) [19]. The AZCERT team rigorously evaluates every drug and constantly updates these evaluations using information available from basic science and clinical evidence from the medical literature, the US FDA summary basis of drug approvals, and labels. For this study, we included drugs from all three risk group lists: Drugs with a known risk of TdP (list 1) are generally accepted as causing TdP; drugs with a possible risk of TdP (list 2) have been associated with TdP in some reports, but compelling evidence for causing this arrhythmia is lacking; drugs with a conditional risk of TdP (list 3) carry a risk of TdP under certain conditions, such as in patients with congenital LQTS, drug overdose, or coadministration of interacting drugs. Our analysis was based on data extracted from the CredibleMeds® website on 23 August 2017. Drugs considered to have no risk of TdP by AZCERT were not considered in the analysis. Since hospitalization in the acute medical units (AMU) in our institution can last ≤ 3 days before admission to internal medicine units, we define chronic use as treatment with a QT-prolonging agent for ≥ 4 days. The IBM Micromedex Drugdex® [20] drug interactions tool was used to evaluate the medicines prescribed for each patient to identify the presence of QT-DDIs. The mechanism of each QT-DDI was classified as either pharmacokinetic or pharmacodynamic. Pharmacokinetic QT-DDIs were defined as attributable to alterations in the metabolism of QT-prolonging medications. Pharmacodynamic QT-DDIs were defined as attributable to the effects of two drugs identified as directly causing prolongation of the QTc interval (mainly blocking the hERG potassium channel).

2.2 Statistical Analysis

Continuous variables were described as mean and standard deviation. Possible statistical associations between QTc interval and continuous variables were analyzed using the bivariate Pearson correlation analysis. Categorical variables were expressed as number and percentage. Patients were divided into two subgroups according to the presence of LQTS, and a case–control analysis was performed. For continuous variables, the two-tailed independent t test or the Mann–Whitney test (in case of non-normal distribution) were used. To explore statistical correlation between LQTS and dichotomous variables, the two-sided χ2 test or the Fisher’s exact test (if n < 5) were used. Variables that were significant in the univariate analysis were then tested in a multivariate model using logistic regression. Statistical significance was set at p < 0.05 for all comparisons. All statistical analyses were performed using SPSS v.24.0 (IBM Corp., Armonk, NY, USA).

3 Results

A total of 243 patients were enrolled in the study: mean age, 79.65 ± 8.27 years, males, n = 121 (40.8%), mean QTc interval, 453.70 ± 43.77 ms. Figure 1 shows the flow chart of the full screening process. Table 1 displays demographic and clinical characteristics and a comparison of variables in patients with and without LQTS on admission. Overall, 89/243 (36.6%) patients had LQTS. A QTc interval prolongation > 500 ms was present in 29/243 (11.9%) patients. A vast majority were prescribed at least one QT-prolonging drug (218/243 [89.7%]), whereas 74/218 (30.5%) were on at least one medication with known risk of TdP (list 1). Notably, more than one-third of our cohort was receiving treatment with more than two QT-prolonging agents at the time of the ECG.

Fig. 1
figure 1

Selection of study sample AV atrioventricular

Full size image
Table 1 Demographic, clinical, and laboratory characteristics of patients and comparison of variables in subjects with and without long QT syndrome on admission
Full size table

Serum potassium, calcium, and magnesium levels were available in 237 (97.5%), 235 (96.7%), and 229 (94.2%) patients, respectively. The most common electrolyte imbalance observed was hypocalcemia (63 [26.8%]). Moreover, a significant proportion of patients had underlying cardiac diseases, more frequently left ventricular hypertrophy (52 [21.4%]), dilated cardiomyopathy/heart failure (43 [17.7%]), and chronic ischemic cardiopathy (34 [14.0%]).

Table 2 shows the number and TdP risk lists of QT-prolonging drugs prescribed in 218 patients. A total of 498 QT-prolonging agents were used at the time of ECG: the most prescribed were furosemide (121/498 prescriptions [24.3%]), pantoprazole (91/498 [18.3%]), and lansoprazole (35/498 [7.0%]). Clarithromycin (19/498 prescriptions [3.8%]), amiodarone (13/498 [2.6%]), and citalopram (11/498 [2.2%]) were the most used medications with a known risk of TdP (list 1).

Table 2 Prescription of QT-prolonging drugs (more than two prescriptions) stratified by risk category and therapeutic classes
Full size table

A total of 31/243 (12.8%) patients presenting with 44 QT-DDIs were identified (Table 3). Multiple QT-DDIs were identified in 9/31 (29.0%) subjects: 6/31 (19.3%) presented with two QT-DDIs, 1/31 (3.2%) with three QT-DDIs, and 2/31 (6.4%) with four QT-DDIs. The most frequent (n > 1) drug interacting pairs were citalopram–omeprazole (4/44 [9.1%]), citalopram–pantoprazole (3/44 [6.8%]), quetiapine–trazodone, citalopram–clarithromycin, clarithromycin–amiodarone, and clarithromycin–paroxetine (2/44 [4.5%]). According to CredibleMeds® classifications, drugs with a known risk of TdP (list 1), a possible risk of TdP (list 2), and a conditional risk of TdP (list 3) were involved in 32/44 (72.7%), 7/44 (15.9%), and 33/44 (75.0%) QT-DDIs, respectively. The most common medications (n > 4) implicated in QT-DDIs were citalopram (n = 8), quetiapine (n = 6), metronidazole (n = 6), and clarithromycin (n = 5). According to Micromedex®, the majority of the QT-DDIs were pharmacodynamic (n = 26), but the two most frequent QT-DDIs (citalopram–omeprazole [n = 4] and citalopram–pantoprazole [n = 3]) were ascribable to a pharmacokinetic mechanism.

Table 3 Two or more QT-prolonging drug–drug interactions according to CredibleMeds® risk category and mechanism
Full size table

Table 4 presents the results of the Pearson correlation for continuous variables. Significant positive associations were present between QTc and age (r = 0.159, p = 0.013) and number of QT interval-prolonging drugs used (r = 0.200, p = 0.002), whereas significant negative associations existed between QTc interval and potassium levels (r = − 0.280, p < 0.0001), calcium levels (r = − 0.166, p = 0.011), and magnesium levels (r = − 0.133, p = 0.044).

Table 4 Pearson correlation coefficients of quantitative determinants of corrected QT interval prolongation
Full size table

Results of the univariate analysis of the dichotomous variables are shown in Table 5. A significant correlation was identified between LQTS and the use of more than one QT interval-prolonging drug (odds ratio [OR] 1.87; 95% confidence interval [CI] 1.06–3.31; p = 0.029), use of more than two QT interval-prolonging drugs (OR 2.18; 95% CI 1.26–3.76; p = 0.005), chronic use of QT interval-prolonging drugs (OR 2.29; 95% CI 1.25–4.18; p = 0.006), male sex (OR 2.16; 95% CI 1.26–3.68; p = 0.004), hypocalcemia (OR 2.06; 95% CI 1.14–3.71; p = 0.015), hypothyroidism (OR 3.84; 95% CI 1.39–10.64; p = 0.006), and dilated cardiomyopathy/heart failure (OR 3.35; 95% CI 1.58–7.09; p = 0.001). Multivariate analysis of the seven significant variables determined by univariate analysis identified five factors significantly associated with LQTS (Table 6): male sex (OR 2.85; 95% CI 1.56–5.22; p = 0.001), hypocalcemia (OR 2.18; 95% CI 1.13–4.20; p = 0.019), hypothyroidism (OR 3.74; 95% CI 1.25–11.25; p = 0.019), dilated cardiomyopathy/heart failure (OR 3.67; 95% CI 1.58–8.51; p = 0.002), and chronic use of QT-prolonging drugs (OR 2.37; 95% CI 1.13–4.99; p = 0.023).

Table 5 Univariate analysis of dichotomy risk factors of long QT syndrome
Full size table
Table 6 Multivariate model by logistic regression of dichotomy risk factors of long QT syndrome
Full size table

4 Discussion

In this study, we observed a high proportion (89/243 [36.6%]) of LQTS in a cohort of consecutive older medical inpatients and, notably, 11.9% (29/243) of QTc interval prolongation > 500 ms. Our prevalence rates were significantly greater than those described in two previous retrospective studies in geriatric inpatients. Using the same definition of LQTS, Lubart et al. [12] reported QTc interval prolongation in 27% (115/422) of patients admitted to an acute geriatric ward (7.15% with QTc interval > 500 ms). In a 6-month study performed in a French university hospital, Maison et al. [5] found LQTS in 22% (33/152) of patients (5% with QTc interval > 500 ms). In studies applying Bazett’s correction, a comparable prevalence of QTc interval > 500 ms of 10.7% was found in 412 patients aged > 18 years consecutively admitted to an intensive care unit [21], and a considerably higher prevalence of 18.2% was reported by Tisdale et al. [22] in patients (n = 900) admitted to cardiac care units. Although it seems plausible that the use of Bazett’s formula with thresholds of 450 ms in men and 470 ms in women may have resulted in an overestimation of overall patients with LQTS in this study, the prevalence of QTc > 500 ms may not be affected by the use of this correction formula, given that the application of Fridericia’s formula to QTc interval in our cohort led to a similar prevalence of 10.3%.

Another relevant finding of this study is that drugs classified in the CredibleMeds® lists were used in the majority of older patients on admission to our internal medicine unit: nearly 90% (218/243) of subjects were receiving at least one QT interval-prolonging drug at the time of ECG. In addition, more than 30% (74/243) of patients were under treatment with agents in CredibleMeds® list 1 (known risk of TdP). We postulated that this finding was because of a large use of QT interval-prolonging medications during the hospitalization. Indeed, 53/243 (21.8%) patients may have had prescriptions of QT interval-prolonging agents during their stay in the emergency department and/or in the AMU. This hypothesis is consistent with several studies reporting a trend towards adding QT interval-prolonging drugs in different hospital settings, either during hospitalization or at discharge. Pasquier et al. [11] found that half of patients with prolonged QT interval at the time of admission received at least one additional QT interval-prolonging drug during the hospital stay in an internal medicine unit. Tisdale et al. [22] also reported that 87/251 (34.7%) patients admitted to cardiac care units with QTc interval prolongation were subsequently administered QT interval-prolonging drugs. Franchi et al. [16] observed the use of QT interval-prolonging drugs in 55.2% of geriatric inpatients on admission that increased to 64.2% at discharge, a finding consistent with that of Schächtele et al. [9], who found that 58.7% of patients were prescribed QT interval-prolonging drugs at discharge from geriatric units. In our cohort, 165/243 (67.9%) patients were chronic users of QT interval-prolonging agents, a finding consistent with that of Maison et al. [5], who reported the use of QT interval-prolonging medications in 61.8% (94/152) of patients at the time of ECG. As we defined chronic use as treatment with a QT interval-prolonging agent for ≥ 4 days, it seems plausible that these data highlight prescriptions by general practitioners, suggesting that drug-induced LQTS may be underestimated in this setting.

Frequent co-prescription of QT interval-prolonging drugs has been reported in the literature for the last two decades. A retrospective cohort study using an outpatient database of one of the largest pharmacy benefit managers in the USA identified a prevalence of QT-DDIs of 9.4% (103,119/1.1 million individuals) in 1999 [23]. Despite this, it appears that software to help prescribers detect QT-DDIs is not commonly adopted and a retrospective database study [24] found that, even with the support of an electronic system alert, ECG monitoring was performed in only 33% of patients prescribed two or more medications associated with LQTS. In the present study, we found that Micromedex®, one of the most reliable and sensitive software programs for the analysis of DDIs [25], may be insufficient to quantify the cumulative pharmacological QT interval-lengthening risk in internal medical units. The systematic use of this tool in our cohort identified QT-DDIs in only 31/243 (12.8%) patients, largely underestimating pharmacodynamic QT-DDIs, given that almost two-thirds of our patients received two or more medications that prolonged the QT interval according to CredibleMeds® lists at the time of ECG.

In line with previously published papers showing an increased risk of LQTS with the use of QT interval-prolonging drug combinations compared with single drugs, we observed that QTc interval correlated with the number of QT interval-prolonging agents at Pearson correlation. In animal studies, the simultaneous administration of two potassium channel blockers markedly potentiated the increase of the QTc interval [26, 27]. Moreover, most clinical studies have highlighted an enhanced TdP risk in patients prescribed multiple QT interval-prolonging drugs. A retrospective case–control study that quantified the risk of cardiac arrest related to the use of non-arrhythmic LQTS inducers in 140 patients resuscitated from cardiac arrest at a Dutch university hospital found an increased risk of arrest associated with the number of prescribed LQTS-associated drugs (OR 4.8; 95% CI 1.6–14) [28]. Drug interactions also represented a high proportion of cases of drug-induced TdP within a Belgium study using the EudraVigilance database (18/31 cases) [29]. However, some evidence in the literature is conflicting. A retrospective study based on a US hospital database that quantified the risk of LQTS in patients receiving medicines classified in CredibleMeds® lists did find QTc interval prolongation (p < 0.001) in patients receiving a single medication with a known risk of TdP compared with those receiving medications with no risk of TdP but found no statistically significant difference in the QTc interval among patients prescribed multiple QT interval-prolonging drugs [30].

As highlighted in several studies, a limited number of agents account for the majority of prescriptions of QT interval-prolonging drugs in geriatric inpatients, namely furosemide, citalopram, and amiodarone. Indeed, in the previously cited German study [9], citalopram and amiodarone were identified as the first and third most commonly used drugs with a known risk of TdP, and only 20 QT-DDIs accounted for more than 90% of co-prescriptions of QT interval-prolonging agents. Thus, it seems plausible that an accurate analysis of a restricted number of drugs could be very relevant in the clinical management of the risk of TdP in older inpatients.

Our finding that furosemide was the most used QT interval-prolonging drug in geriatric inpatients is in agreement with findings of other recent studies [5, 16]. The association between diuretics and in-hospital TdP may be explained by its correlation with congestive heart failure, hypokalemia, and hypomagnesemia [1]. Although the use of loop diuretics was an independent risk factor for QTc interval prolongation [6, 31], the risk of associated TdP is generally considered low. However, as furosemide is the mainstay of treatment of severe cardiovascular diseases, greater attention should be paid to frequently co-prescribed QT interval-prolonging medications.

In our cohort, proton pump inhibitors (PPIs) were the second most commonly prescribed class of QT interval-prolonging drugs, a finding consistent with the Italian REPOSI study [16], which found pantoprazole to be the second most frequently prescribed QT interval-prolonging agent on admission (340/2156 [11.8%]). We found a total of 154/498 (25.1%) PPI prescriptions, and PPIs were frequently co-administered with furosemide (n = 151) and amiodarone (n  = 19). Despite accumulating evidence, PPIs do not currently receive attention as a factor potentially contributing to QTc interval prolongation in clinical practice. In a very recent Swedish prospective register-based cohort study on the occurrence of TdP in relation to drugs listed in the CredibleMeds® classification, 131/410 (32%) patients with TdP were also receiving PPIs [32]. Although rare, PPI-induced hypomagnesemia can be associated with TdP [33] according to experimental studies showing that cytosolic magnesium promotes repolarization of myocardial cells via its modulating effects on several potassium channels, including IKr and the transient outward current [34]. Nevertheless, PPIs may also significantly influence the QTc, regardless of hypomagnesemia. In this study, we identified five cases of LQTS (three with QTc interval > 500 ms) in eight patients co-administered ceftriaxone and lansoprazole. This finding is in agreement with recent evidence showing that lansoprazole in combination with ceftriaxone carries a risk of QTc interval prolongation, as it has been demonstrated that ceftriaxone + lansoprazole can significantly block the hERG channel and inhibit IKr (− 57.6%) [35, 36]. According to these data, and given that PPIs are often prescribed inappropriately [37, 38], the use of this class of drugs in older people who present with a cumulative risk of TdP should be accurately evaluated either during hospitalization or at hospital discharge.

Citalopram and sertraline were the most prescribed antidepressants in our cohort. Citalopram was implicated in more QT-DDIs (n = 12; 4.94% of the study sample), notably eight pharmacokinetic QT-DDIs (including in combination with omeprazole in four patients, who might therefore have been exposed to significantly higher citalopram concentrations (+ 35.3%) because of omeprazole’s inhibition of cytochrome P450-2C19 [39]) and four pharmacodynamic QT-DDIs (see the electronic supplementary material). Despite general agreement in the literature, including a recent meta-analysis [40], that citalopram is the selective serotonin reuptake inhibitor more consistently associated with QTc prolongation, and recommendations from the FDA [41] and other regulatory authorities, awareness of this issue in older patients seems to be lacking.

The multivariate analysis identified non-pharmacological risk factors associated with LQTS in the literature, including hypocalcemia, hypothyroidism, and dilated cardiomyopathy/heart failure. However, in our opinion, the most relevant finding was the independent association between male sex and LQTS (OR 2.83; 95% CI 1.55−5.19; p = 0.001). Although females are generally considered at greater risk for drug-induced TdP than males [1], this surprising correlation has also been observed in three previous studies in older inpatients [5]. In a cross-sectional study in 178 residents of long-term care wards in a geriatric hospital, Lubart et al. [42] found a statistically significant prevalence of prolonged QTc interval in males versus females, whereas the previously cited paper by Maison et al. [5] reported an association between male sex and QTc interval prolongation after multivariate analysis (OR 3.25; 95% CI 1.43–7.41; p = 0.004). More recently, in an Italian study in 432 patients (mean age 83.3 years), Bo et al. [6] observed a significant association between male sex and QTc interval prolongation (OR 2.09; 95% CI 1.34–3.26). Physiological evidence could at least partly explain this finding. The mean QTc interval is longer in women than in men from adolescence until the sixth decade, but this difference tends to disappear in very old patients. Recently, Rautaharju et al. [43] demonstrated that, while the QTc interval still increases with increasing age in the elderly, the slope appears to be greater in men, and the men’s curve almost crosses the women’s curve in patients aged over 80 years, highlighting the fact that the sex difference in QTc may arise from a shortened QT interval in adolescent males, with no shortening in adolescent females, as documented in a previous study of the same group [44]. The shorter QTc interval in men is likely due to the testosterone-associated altered dynamics of cardiac ion channels [45, 46]: preclinical studies showed that testosterone can both increase the repolarizing potassium currents IKr and IKs and decrease the depolarizing L-type calcium current, ICaL [45]. Furthermore, Bidoggia et al. [47] found that the QT interval was prolonged in castrated compared with non-castrated men and that women with hyperandrogenism had shorter QT intervals than controls. Zhang et al. [48] compared multivariate-adjusted QT intervals in the highest and lowest quartiles of total and free testosterone levels in men in the Third National Health and Nutrition Examination Survey and in men and postmenopausal women in the Multi-Ethnic Study of Atherosclerosis. They identified significant differences in the adjusted QT interval between the highest and lowest quartiles of testosterone in men but not in women [48]. In addition, in recent years, evidence has supported the hypothesis that androgen-deprivation therapy (ADT) can increase the risk of TdP [49]. Clinical interventional studies in men with prostate cancer demonstrated that ADT is associated with QTc prolongation [50,51,52,53]. In a prospective cohort of 58 consecutive unselected patients experiencing TdP, ADT unexpectedly represented the second most frequently administered QT interval-prolonging drug in males (3/19 [16%]), after amiodarone [54]. Moreover, Salem et al. [55] reported that the androgen receptor antagonist enzalutamide significantly prolonged action potential duration in induced pluripotent stem cell-derived cardiomyocytes. These effects were associated with IKr inhibition and INa enhancement and were reversed by dihydrotestosterone. However, on the basis of the available evidence, a similar risk of LQTS should be identified in very old men and women, so further research is needed to clarify this issue.

4.1 Study Limitations

The present study has several limitations. First, this study enrolled patients admitted to a single medical unit. Second, this was a cross-sectional design without follow-up of possible arrhythmic events in patients with LQTS on admission. Other limitations are related to the use of Bazett’s formula, as it tends to overcorrect QT intervals at a faster HR; moreover, we used the clinical standards of 450 ms in men and 470 ms in women, which may have overestimated the number of patients with prolonged QT interval when using QTcB. In addition, the high proportion of patients who did not provide informed consent and so were excluded from the study, may have affected the true prevalence of LQTS in our cohort. Recall bias is also possible as participants were asked to remember information about their prescriptions from their general practitioners on admission.

5 Conclusions

The prevalence of LQTS with QTc interval > 500 ms in geriatric inpatients was > 10%, and QT interval-prolonging drugs were frequently used on admission (more than 30% of patients were treated with drugs with a known risk of TdP). Male sex was independently associated with LQTS in our cohort. PPIs were the second most frequently prescribed class of drugs, after diuretics. Given the increasing evidence on the risk of LQTS from PPIs, education on the appropriate prescription of this class of drugs in older patients, who are at particular risk for life-threatening arrhythmias, is needed. Further larger population-based studies are warranted to better define the role of PPIs and sex as a risk factor of LQTS in very old patients.