Thiazide diuretics use & risk of falls & syncope among hypertensives: A retrospective cohort study

Materials & Methods

Study design and setting

This retrospective cohort study was conducted at the department of General Medicine, Guru Gobind Singh Government Hospital and Shri M. P. Shah Government Medical College, Jamnagar, Gujarat between March 2019 and March 2024. The Institutional Ethics Committee examined and approved the study protocol.

Eligible population

All adult individuals aged ≥18 yr with a prescription of thiazide diuretic agents for the treatment of hypertension within the study period were eligible. Subjects were identified from the EMR system with the help of relevant prescription and billing codes. Individuals with incomplete medical records or who were lost to follow up were excluded from the study.

Sample size calculation

In this study, we assumed a thiazide diuretic use and falls/syncope events association of a moderate effect size-odds ratio of 1.5, based on previous literature. For these estimated parameters and considering a desired power of 80 per cent at a two-sided alpha level of 0.05, we estimated a required sample size of 472 (236 in each group)¹⁰.

Sampling strategy

Consecutive inclusion of the study participants receiving thiazide diuretic medication who were enrolled for hypertension was considered from March 2019 to March 2024 at the Guru Gobind Singh Government Hospital and Shri M P Shah Government Medical College, Jamnagar, Gujarat. This was done to reduce the possibility of selection bias and guarantee that the sample was representative of the target population (individuals with hypertension on thiazide diuretics). The steps performed for this sampling technique are described in the figure.

Close

Figure.

Steps for sampling strategy.

Export to PPT

Results

The study included 236 study participants taking thiazide diuretics and 236 study participants not taking thiazides (Table I). The mean age of study participants in the thiazide group (68.4±9.6 yr) was similar to the non-thiazide group (67.8±10.2 yr), with no significant difference (P=0.542). The gender distribution was balanced, with 50 per cent males in both groups. The mean duration of thiazide use in the thiazide group was 22.8±15.2 months. Study participants in the non-thiazide group, by definition, had no thiazide use. The mean duration of hypertension was similar between the two groups (thiazide group: 9.6±6.4 yr; non-thiazide group: 8.8±5.6 yr, P=0.168). Diabetes mellitus (7.2±4.6 yr vs. 6.8±4.2 yr, P=0.336) and congestive heart failure (5.4±3.2 yr vs. 4.8±2.8 yr, P=0.042) were slightly higher in the thiazide group, with a statistically significant difference observed for congestive heart failure. Falls/syncope events were more in the thiazide group (32.2%) compared to the non-thiazide group (19.5%); P=0.002 .

Study participants taking thiazides had a higher prevalence of electrolyte disturbances compared to that in controls (Table II). Hyponatremia was observed in 57 (24.1%) study participants in the thiazide group versus 31 (13.1%) in the non-thiazide group (P=0.04). Hypokalaemia was present in 52 (22%) thiazide users compared to 32 (13.6%) non-users (P=0.02). Hypercalcemia was found in 37 (15.7%) study participants taking thiazides versus 17 (7.2%) not taking thiazides (P=0.006). Metabolic alkalosis was significantly more common in the thiazide group, affecting 89 (37.7%) study participants, compared to 52 (22%) in the non-thiazide group (P<0.001).

The severity of hyponatremia also differed between groups. In the thiazide group, 35 (61.4%) study participants had mild hyponatremia, 19 (33.3%) had moderate hyponatremia, and three (5.3%) had severe hyponatremia. In contrast, in the non-thiazide group, 20 (64.5%) had mild, eight (25.8%) had moderate, and three (9.7%) had severe hyponatremia.

Similarly, for hypokalaemia, in the thiazide group, 32 (61.5%) study participants had mild, 12 (23.1%) had moderate, and eight (15.4%) had severe hypokalaemia. In the non-thiazide group, 16 (50%) had mild, 11 (34.4%) had moderate, and five (15.6%) had severe hypokalaemia. The prevalence of hypercalcemia was significantly higher in the thiazide group (15.7%) compared to the non-thiazide group (7.2%); P=0.006. The prevalence of metabolic alkalosis was also significantly higher in the thiazide group (37.7%) compared to the non-thiazide group (22%), with a P<0.001.

Supplementary table explains the renal function and markers.

Table III presents that increasing age was associated with falls/syncope in the thiazide group. For every one-year increase in age, the adjusted odds of falls/syncope increased by 38 per cent [adjusted odds ratio (AOR)=1.38, 95% confidence interval (CI): 1.14-1.67, P=0.001]. Study participants with congestive heart failure had a significantly higher risk of falls/syncope while taking thiazides, with an adjusted odds ratio of 2.12 (95% CI: 1.28-3.51, P=0.003) compared to those without congestive heart failure.

Electrolyte disturbances, such as hyponatremia and hypokalaemia, were associated with falls/syncope in the thiazide group. The presence of metabolic alkalosis was also a significant risk factor. Study participants with (AKI) had an increased risk of falls/syncope while taking thiazides, with an AOR of 1.85 (95% CI: 1.04-3.29, P=0.036) compared to those without AKI. The estimated glomerular filtration rate (eGFR) was inversely associated with the risk of falls/syncope (AOR =1.22, 1.07-1.39, P=0.003).

Although CKD and hypercalcemia had significant crude odds ratios (CORs) for falls/syncope, their associations were not statistically significant after adjusting for other factors in the multivariate logistic regression analysis. Overall, this logistic regression analysis identified several risk factors, including age, congestive heart failure, electrolyte disturbances (hyponatremia, hypokalaemia, and metabolic alkalosis), AKI, and decreased renal function, that were independently associated with an increased risk of falls/syncope in study participants taking thiazides.

Table IV presents a comprehensive analysis of the association between thiazide diuretic use, its duration, and the risk of falls/syncope events in hypertensive individuals. The table has two main sections: overall thiazide use and duration. The first section shows that among study participants not taking thiazides, 46 out of 236 (19.5%) experienced falls or syncope, while in the thiazide group, this increased to 76 out of 236 (32.2%). After adjusting for potential confounders, the association remained significant, with an AOR of 1.85 (95% CI: 1.18-2.9).

The second section of the table presents the impact of thiazide use duration on falls/syncope risk. A clear dose-response relationship is evident, with longer durations of thiazide use associated with progressively higher risks. Compared to no thiazide use, study participants using thiazides for less than 12 months showed a non-significant increase in risk (AOR: 1.48, 95% CI: 0.81-2.7). However, for those using thiazides for 12-24 months and more than 24 months, the risk became statistically significant, with AORs of 1.92 (95% CI: 1.07-3.45) and 2.15 (95% CI: 1.21-3.83), respectively.

Discussion

The present study has identified the potential risks related to thiazide diuretic use in hypertensive individuals. This observation is consistent with previous studies reporting an elevated risk of falls and syncope related to thiazide diuretic use in the general population^8,9.

Firstly, the higher prevalence of electrolyte disturbances, such as hyponatremia, hypokalaemia, and hypercalcemia, among study participants taking thiazide diuretics is a well-established adverse effect of these medications^17-19. These electrolyte imbalances can lead to various physiological consequences, including neurological impairments, muscle weakness, and cardiovascular dysregulation, which may predispose patients to falls and syncope events^20,21.

Additionally, we observed a higher prevalence of metabolic alkalosis in the thiazide group, another known adverse effect of these diuretics⁷. Metabolic alkalosis can contribute to impaired neuromuscular function, altered mental status, and increased risk of arrhythmias, further exacerbating the risk of falls and syncope²².

The study also revealed a higher prevalence of AKI and CKD among study participants taking thiazide diuretics. Renal impairment can lead to various metabolic and electrolyte disturbances and impaired fluid and electrolyte homeostasis, which may contribute to an increased risk of falls and syncope^23,24.

Furthermore, the logistic regression analysis identified several independent risk factors for falls and syncope in study participants taking thiazide diuretics. These included increasing age, congestive heart failure, longer duration of thiazide use, hyponatremia, hypokalaemia, metabolic alkalosis, AKI, and decreased estimated glomerular filtration rate (eGFR). These findings highlight the importance of considering individual patient characteristics, comorbidities, and potential drug-related adverse effects when prescribing thiazide diuretics, particularly in vulnerable patient populations.

Our findings of increased risk of falls and syncope associated with thiazide diuretic use are consistent with several previous studies. The present study also provided new insights into potential mechanisms. A large retrospective cohort study by Ravioli et al¹⁰ (2021) involving 1604 patients found such an association similar to the AOR of 1.96 (95% CI 1.28-3.01) for falls/syncope events that we estimated.

A meta-analysis by Bai et al²⁵ (2023) of 13 studies that were meta-analysed found the same. Our findings of increased risk associated with hyponatremia and hypokalaemia are novel contributions to the literature, as previous studies have not specifically examined these electrolyte abnormalities related tofall- risk in thiazide users.

Interestingly, prospective research by Bakar et al²⁶ (2021) discovered that older persons using diuretics (AOR 2.803, 95% CI 1.418, 5.544) and polypharmacy (AOR 2.513, 95% CI 1.339, 4.718) had a greater risk of falling. However, no correlation was found between an increased occurrence of falls and older persons, especially those taking more anti-hypertensive drugs²⁶.

A prior meta-analysis found that taking thiazides was linked to a lower incidence of fracture in case-control studies but not in cohort studies. Selection and recollection biases are two types of inherent biases that may be responsible for the relationships shown in case-control studies. Therefore, using thiazides might not prevent fractures²⁷. While our study demonstrated associations between thiazide use and falls/syncope, several alternative explanations should be considered. Confounding by indication may play a role, as patients on thiazides could have more severe hypertension or other comorbidities that independently increase the risk of falls. Additionally, unmeasured confounders such as cognitive function, frailty, or physical activity levels could influence thiazide prescription and the risk of falls. The potential for detection bias should also be acknowledged, as study participants taking thiazides may be monitored more closely, leading to increased identification of falls/syncope events.

Future research should focus on prospective studies to further elucidate the causal relationship between thiazide diuretic use and falls and syncope events and the underlying mechanisms involved. Additionally, studies investigating the efficacy of preventive interventions and alternative treatment strategies in high-risk patients would be valuable in improving patient safety and clinical outcomes.

The present study included 152 study participants taking thiazide diuretics and 352 study participants not taking thiazides, providing sufficient statistical power to detect significant associations. Moreover, in this study, detailed demographic, clinical, laboratory, and outcome data were collected, allowing for a thorough investigation of potential risk factors and confounders. The use of propensity score matching in the control group further helped reduce potential selection bias and ensure comparability between the two groups. The inclusion of multivariate logistic regression models allowed for the adjustment of potential confounding factors, providing more reliable estimates of the associations between thiazide use and the risk of falls/syncope.

Despite the above strengths, there were some limitations. As a retrospective study, the data were collected from existing medical records in the present investigation, which may be subject to potential inaccuracies or missing information. The investigation was located in one centre study. Notably, the study did not provide details on the severity or consequences of the falls and syncope events, which could have been useful for assessing their clinical impact. Despite the comprehensive data collection and multivariate analysis, there may be additional, unmeasured confounding factors that could influence the observed associations. The unmeasured confounders like diet, physical activity, depression, and cognitive impairment were not accounted for and might have influenced the results. Finally, the study did not specify the duration of follow up, which could affect the capture of long-term outcomes or adverse events.

Overall, our study demonstrated a significant increase in falls and syncope events among patients taking thiazide diuretics compared to those who were not on these medications. We identified several independent risk factors, including increasing age, congestive heart failure, longer duration of thiazide use, electrolyte imbalances (hyponatremia, hypokalaemia, and metabolic alkalosis), acute kidney injury, and decreased renal function. These results highlighted the importance of careful monitoring and individualized treatment approaches when prescribing thiazide diuretics to hypertensive patients. While these medications remain valuable for hypertension management, our findings underscore the need for regular electrolyte and renal function monitoring and the implementation of fall prevention strategies in high-risk patients.