Upper Urinary Tract
Urinary calculi and an increased risk of stroke: a population-based follow-up study
Herng-Ching Lin, School of Health Care Administration, Taipei Medical University, 250 Wu-Hsing St., Taipei 110, Taiwan. e-mail: email@example.com
Study Type – Prognosis (cohort)
Level of Evidence 2b
What's known on the subject? and What does the study add?
Although early studies failed to detect an association between urinary calculi (UC) and subsequent cardiovascular risk, there is growing evidence among more recent research supporting this association with some studies more specifically suggesting that stroke is a major concern for UC sufferers. After adjusting for potential confounding factors, UC patients were more likely to have experienced a stroke then those without UC during the five-year follow-up period (hazard ratio = 1.43, 95% Cl = 1.35–1.50, P < 0.001).
- • To examine in a population-based study the relationship between a history of nephrolithiasis and/or ureterolithiasis and the subsequent risk of stroke, as previous studies have shown that stone disease is associated with several cardiovascular risk factors. However, none of the studies that have investigated the relationship between urinary calculi (UC) and stroke were able to detect an association at a significant level.
PATIENTS AND METHODS
- • We used data sourced from the Taiwan Longitudinal Health Insurance Database 2000.
- • In all, 25 181 adult patients newly diagnosed with UC were recruited as a study cohort, along with 125 905 matched enrolees with no history of stone disease as a comparison cohort.
- • All the subjects were tracked for a 5-year period beginning from their index ambulatory care visits, and those who subsequently had a stroke identified.
- • Cox proportional hazards regressions were used to compare the risk of stroke between the study and comparison cohorts.
- • During the 5-year follow-up period, the incidence rate of stroke was 1.78 (95% confidence interval [CI] 1.71–1.86) per 100 person-years in patients with UC and 1.25 (95% CI 1.22–1.27) per 100 person-years in patients without UC.
- • After adjusting for hypertension, diabetes, hyperlipidaemia, cardiovascular disease, urbanization level, gout, and obesity, patients with UC were more likely to have had a stroke than those without UC during the 5-year follow-up period (hazard ratio 1.43, 95% CI 1.35–1.50, P < 0.001).
- • Our results suggest that there is an increased risk of stroke during the first 5 years after a diagnosis of UC.
Longitudinal Health Insurance Database 2000
National Health Insurance
New Taiwan dollar
Urinary calculi (UC) is a common genitourinary disorder with a worldwide lifetime incidence of 10–15% . With the exception of the two World Wars, the incidence of UC has been increasing among both adults and children over the past 100 years [2–4]. Therefore, considering the relatively high and increasing incidence rate of UC, it is important to understand what sequelae may affect the many survivors of this low-mortality condition.
Although early studies failed to detect an association between UC and subsequent cardiovascular risk [5,6], there is growing evidence among more recent research supporting this association [7–11], with some studies more specifically suggesting that stroke is a major concern for patients with UC [12,13]. Stroke is a major cause of mortality and morbidity worldwide and is a significant cause of disability in adults with substantial economic consequences. Furthermore, preventative measures can be taken to reduce the risk of stroke, e.g. blood pressure control and the use of statins and aspirin . However, no published study to date investigating the relationship between UC and stroke has been able to detect an association at a significant level. Therefore, using a population-based data set from Taiwan, the present study sought to investigate the relationship between UC and the subsequent risk of stroke during a 5-year follow-up period.
PATIENTS AND METHODS
The data analysed in this study was sourced from the Longitudinal Health Insurance Database 2000 (LHID2000), which is derived from the Taiwan National Health Insurance (NHI) programme. This single payer programme was launched on 1 March 1995 to provide affordable and comprehensive medical care to all of Taiwan's residents. By 2007, the programme covered 22.60 million of Taiwan's 22.96 million residents (coverage rate 98.4%). As the vast amount of data concerning all of the enrolees under the NHI would be cumbersome to analyse, the Taiwan National Health Research Institute produced and released the LHID2000 to aid researchers in Taiwan. The LHID2000 contains the registration files and all the original claims data of 1 million individuals systematically and randomly sampled from the 2000 Registry for Beneficiaries of the NHI programme. The Taiwan National Health Research Institute claims that there are no significant differences in the gender distribution between the enrolees in the LHID2000 and all enrolees under the NHI programme. Many researchers have sourced and analysed data from the LHID2000 to conduct and publish a range of studies [15–17].
After consulting with the director of the Taipei Medical University Institutional Review Board, this study was exempted from full review because the LHID2000 consists of de-identified secondary data released to the public for research purposes.
This prospective study included a study cohort and a comparison cohort. In all, 40 019 patients were identified who had received a diagnosis of UC (ICD-9-CM codes 592 [calculus of kidney and ureter], 592.0 [calculus of kidney], 592.1 [calculus of ureter] or 592.9 [urinary calculus, unspecified]) during ambulatory care visits (including outpatients departments of hospitals and clinics) between 1 January 2001 and 31 December 2003. Their first diagnosis of UC during an ambulatory care visit between 2001 and 2003 was assigned as their index date. Patients who had a diagnosis of UC before their index date (n= 12 921) were then excluded to increase the possibility of only including newly diagnosed cases. Patients aged < 18 years (n= 445) were also excluded to restrict the study to an adult population. Furthermore, all the patients who had received a diagnosis of stroke (ICD-9-CM codes 430–438) before their index ambulatory care visit (n= 1472) were excluded from the assessment. Ultimately, 25 181 patients with UC were included in the study cohort.
The comparison cohort was selected from the remaining enrolees in the registry of beneficiaries of the LHID2000. Similarly, all the enrolees aged < 18 years were excluded. Five controls for every patient with UC (125 905 subjects), matched in terms of sex, age group (<35, 35–39, 40–44, 45–49, 50–54, 55–59, 60–64, 65–69, 70–74, and >74 years), monthly income (in New Taiwan dollars (NT$) 0–15 840, 15 841–25 000, ≥25 001), urbanization level, and the year of index date, were randomly extracted from the LHID2000 data. Their first ambulatory care visit occurring in the index year was assigned as their index date. No selected subjects had ever received a diagnosis of stroke before their index date. However, patients who had been diagnosed with stroke before 1996 may have missed exclusion because the NHI, and therefore the LHID2000, was initiated in 1995. No selected subjects had ever received a diagnosis of UC during the period 1996–2008.
As a result, 151 086 subjects were included in this study. Each patient was individually tracked for a 5-year period starting from their index date to identify the patients who had had any type of stroke (ICD-9-CM codes 430–438). To avoid mistaken diagnoses, we only selected patients who had received at least two diagnoses of stroke from two separate diagnostic teams during the study period. Furthermore, virtually all of the hospitals in Taiwan capable of admitting stroke patients are equipped with CT or MRI scanners, which considerably increases the validity of stroke diagnoses.
Pearson chi-squared tests were used to examine the differences in sociodemographic characteristics and selected co-morbidities between patients with and with no UC. The selected co-morbidities included hypertension, diabetes, coronary heart disease, heart failure, atrial fibrillation, and hyperlipidaemia. All of these have been reported to be risk factors for stroke. These co-morbidities were only included if they occurred in an inpatient setting or in two or more ambulatory care claims coded within 6 months before and after the index ambulatory care visit. We also adjusted for obesity and alcohol abuse/alcohol dependence syndrome in the regression model.
Cox proportional hazards regressions (stratified by sex, age group, monthly income, urbanization level, and the year of index date) were analysed to test the association of UC with subsequent stroke during the 5-year follow-up period, with cases being censored if they died from non-stroke causes during that time. During the study period, 14 882 patients died from non-stroke causes, comprising 2543 from the study cohort (10.1% of the patients with UC) and 12 339 from the comparison cohort (9.8% of the patients with no UC). Sociodemographic characteristics and selected co-morbidities were adjusted for in regression modelling. Hazard ratios (HR) and 95% CIs were computed to present the risk of stroke using a significance level of 0.05.
The distributions of sociodemographic characteristics and selected co-morbidities for the sampled patients are given in Table 1. Of the 151 086 sampled patients, the mean (sd) age was 44.3 (15.7) years. Only 12.2% were aged ≥65 years. Most of the sampled patients were male (62%). After matching for sex, age group, monthly income, and the year of index date, patients with UC had a greater tendency to have hypertension (P < 0.001), diabetes (P < 0.001), coronary heart disease (P < 0.001), heart failure (P= 0.003) and hyperlipidaemia (P < 0.001) than patients with no UC. There were no clinically significant differences in atrial fibrillation between the two cohorts.
Table 1. Demographic characteristics for the sampled patients in Taiwan, stratified by the presence/absence of UC, 2001–2003 (N= 151 086)
|Number of patients||25 181||125 905|| |
|Sex:|| || ||1.000|
| Male||15 618 (62.0)||75 840 (62.0)|| |
| Female||9 563 (38.0)||47 815 (38.0)|| |
|Age group, years:|| || ||1.000|
| <35||6 147 (24.4)||30 735 (24.4)|| |
| 35–39||3 135 (12.4)||15 675 (12.4)|| |
| 40–44||3 386 (13.5)||16 930 (13.5)|| |
| 45–49||3 301 (13.1)||16 505 (13.1)|| |
| 50–54||2 651 (10.5)||13 255 (10.5)|| |
| 55–59||1 801 (7.2)||9 005 (7.2)|| |
| 60–64||1 682 (6.7)||8 410 (6.7)|| |
| 65–69||1 263 (5.0)||6 315 (5.0)|| |
| 70–74||983 (3.9)||4 915 (3.9)|| |
| >74||832 (3.3)||4 160 (3.3)|| |
|Urbanization level:|| || ||1.000|
| 1||7 420 (29.5)||37 100 (29.5)|| |
| 2||7 323 (29.1)||36 615 (29.1)|| |
| 3||4 447 (17.7)||22 235 (17.7)|| |
| 4||3 351 (13.3)||16 755 (13.3)|| |
| 5||2 640 (10.5)||13 200 (10.5)|| |
|Monthly income, NT$:|| || ||1.000|
| 0–15 840||8 383 (33.3)||41 915 (33.3)|| |
| 15 841–25 000||10 714 (42.6)||53 570 (42.6)|| |
| ≥25 001||6 084 (24.2)||30 420 (24.2)|| |
|Hyperlipidaemia||2 666 (10.6)||11 080 (8.8)||<0.001|
|Hypertension||5 059 (20.1)||19 097 (15.2)||<0.001|
|Diabetes||2 306 (9.2)||9 170 (7.3)||<0.001|
|Coronary heart disease||1 805 (7.2)||6 687 (5.3)||<0.001|
|Atrial fibrillation||80 (0.3)||337 (0.3)||0.167|
|Heart failure||385 (1.5)||1 641 (1.3)||0.005|
|Obesity||228 (0.9)||712 (0.6)||<0.001|
|Gout||386 (1.5)||1 155 (0.9)||<0.001|
Table 2 shows the stroke occurrence according to demographic characteristics and co-morbidities. Stroke occurrence was significantly associated with all the included demographic characteristics and selected co-morbidities.
Table 2. Relationships between stroke occurrence and demographic characteristics and co-morbidities for the sampled patients in Taiwan (N= 151 086)
| Male||5765 (6.2)|| |
| Female||3853 (6.7)|| |
|Age group, years:|| ||<0.001|
| <35||322 (0.9)|| |
| 35–39||372 (2.0)|| |
| 40–44||569 (5.9)|| |
| 45–49||940 (9.8)|| |
| 50–54||996 (10.4)|| |
| 55–59||1084 (11.3)|| |
| 60–64||1324 (13.8)|| |
| 65–69||1466 (15.2)|| |
| 70–74||1289 (13.4)|| |
| >74||1256 (13.1)|| |
|Urbanization level:|| ||<0.001|
| 1||2492 (5.6)|| |
| 2||2848 (6.5)|| |
| 3||1480 (5.5)|| |
| 4||1400 (7.0)|| |
| 5||1398 (8.8)|| |
|Monthly income, NT$|| ||<0.001|
| 0–15 840||3185 (6.3)|| |
| 15 841–25 000||4658 (7.2)|| |
| ≥25 001||1775 (4.9)|| |
| Yes||1568 (12.2)|| |
| No||7950 (5.8)|| |
|Coronary heart disease:|| ||<0.001|
| Yes||1737 (18.1)|| |
| No||7881 (5.5)|| |
| Yes||4444 (18.4)|| |
| No||5174 (4.1)|| |
| Yes||1949 (17.0)|| |
| No||7669 (5.5)|| |
| Yes||81 (8.6)|| |
| No||9537|| |
| Yes||214 (13.9)|| |
| No||9404 (6.3)|| |
|Atrial fibrillation:|| ||<0.001|
| Yes||110 (26.4)|| |
| No||9508 (6.3)|| |
|Heart failure:|| ||<0.001|
| Yes||468 (3.1)|| |
| No||9150 (6.1)|| |
Table 3 shows the incidence of stroke during the 5-year follow-up period according to the presence or absence of UC. Of the 151 086 sampled patients, 9618 patients (6.37%) had a stroke during the 5-year follow-up period; 2164 (8.59% of patients with UC) from the study cohort and 7454 (5.92% of patients with no UC) from the comparison cohort. The incidence rate of stroke was 1.78 (95% CI 1.71–1.86) per 100 person-years in patients with UC and 1.25 (95% CI 1.22–1.27) per 100 person-years in patients with no UC. Correspondingly, the log-rank test showed that patients with UC had a significantly lower 5-year stroke-free survival rate than patients with no UC (chi-square value 145.455; P < 0.001).
Table 3. Crude and covariate-adjusted HRs for stroke among the sampled patients during the 5-year follow-up period starting from the index ambulatory care visit
|5-year follow-up period, n (%):|| || || |
| Yes||9 618 (6.37)||2 164 (8.59)||7 454 (5.92)|
| No||141 468 (93.63)||23 017 (91.41)||118 451 (94.08)|
|Incidence rate per 100 person-years (95% CI)||1.34 (1.31–1.37)||1.78 (1.71–1.86)||1.25 (1.22–1.27)|
|Crude HR (95% CI)||–||1.48*** (1.40–1.55)||1.00|
|Adjusted† HR (95% CI)|| ||1.43*** (1.35–1.50)||1.00|
Of the total 9618 sampled patients who had a stroke during the 5-year follow-up period, the mean (sd) number of days between the index date and the occurrence of stroke was 846 (536) days. In patients with UC this figure was 856 days, and in patients with no UC 842 days. There were no significant differences in the period of time between the index date and the onset of stroke between the cohorts (P= 0.186). In addition, there were no significant differences in the type of stroke between the cohorts (P= 0.142).
Table 3 also reports the HRs for stroke by cohort. After adjusting for the patients' urbanization level, hypertension, diabetes, coronary heart disease, heart failure, atrial fibrillation, hyperlipidaemia, obesity, and gout, the stratified Cox proportional analysis results showed that patients with UC were more likely to have had a stroke than those with no UC during the 5-year follow-up period (HR 1.43, 95% CI 1.35–1.50, P < 0.001). Cases were eliminated if individuals had non-stroke mortality during the follow-up period. All the selected co-morbidities including hypertension, diabetes, coronary heart disease, heart failure, atrial fibrillation, and hyperlipidaemia were significantly related to the occurrence of stroke during the 5-year follow-up.
Table 4 further presents the HRs for stroke according to stone location. Among patients with UC, the incidence rate of stroke was 2.07 (95% CI 1.93–2.22) per 100 person-years in patients with kidney calculi, 1.64 (95% CI 1.49–1.79) per 100 person-years in patients with ureter calculi, and 1.71 (95% CI 1.60–1.82) per 100 person-years in patients with unspecified calculi. As compared with subjects in the comparison cohort, the adjusted HR for stroke during the 5-year follow-up period was 1.50 (95% CI 1.38–1.63), 1.42 (95% CI 1.28–1.57), and 1.37 (95% CI 1.27–1.48) for patients with kidney, ureter, with unspecified calculi, respectively.
Table 4. Crude and covariate-adjusted HRs for stroke among the sampled patients according to stone location during the 5-year follow-up period starting from the index ambulatory care visit
|5-year follow-up period, n (%):|| || || || |
| yes||7454 (5.92)||806 (9.84)||465 (7.79)||893 (8.10)|
|Incidence rate per 100 person-years (95% CI)||1.25 (1.22–1.27)||2.07 (1.93–2.22)||1.64 (1.49–1.79)||1.71 (1.60–1.82)|
|Crude HR (95% CI)||1.00||1.73*** (1.61–1.87)||1.34*** (1.22–1.48)||1.40*** (1.30–1.51)|
|Adjusted† HR (95% CI)||1.00||1.50*** (1.38–1.63)||1.42*** (1.28–1.57)||1.37*** (1.27–1.48)|
The is the first prospective population-based study to detect a significantly increased risk of stroke among patients with UC located in the kidney and ureter during a 5-year follow-up period after adjusting for patient demographic characteristics and co-morbid medical disorders. In the present study conducted on 25 181 patients diagnosed with nephrolithiasis, ureterolithiasis, and unspecified stones of the kidney and ureter, we found that the patients of all three UC diagnoses had significantly higher risks of subsequent stroke than comparison patients during a 5-year follow-up period. After adjusting for age, gender, geographic region, and co-morbid medical disorders (hypertension, diabetes, hyperlipidaemia, cardiovascular disease, gout, and obesity), both nephrolithiasis and ureterolithiasis remained significant predictors of subsequent stroke with the combined patients with UC being 1.43-times more likely than the comparison group to have a stroke within the following 5 years. After further stratifying by stone location, we found patients diagnosed with nephrolithiasis, ureterolithiasis, and both nephrolithiasis and ureterolithiasis to be at 1.50-, 1.42-, and 1.37-times the risk of comparison patients of having a stroke during the follow-up period.
The present study found an increased risk for stroke among patients with UC; however, we failed to detect a difference in the time between the index date and stroke between the two groups. As we showed UC to be an independent risk factor for stroke, it might be expected to occur after a shorter follow-up period. But, the results of the present study are in accordance with those of a similar previous study. That study also failed to detect a difference in the time between the index date and stroke onset between a study group of hip replacement patients and a matched comparison group .
Although earlier studies failed to detect an association between UC and subsequent cardiovascular risk [5,6], there is growing evidence supporting this association [7–11], with one recent study clearly showing an increased risk of myocardial infarction among patients with nephrolithiasis . While both myocardial infarction and stroke are vascular events, only two studies to date have explored the association between stroke and nephrolithiasis [12,13].
The first study to investigate the association between stroke and nephrolithiasis was performed by Li et al.  on a population-based sample of normotensive subjects in Sweden. They found that significantly more stroke victims included in their study had a history of renal calculus (3.6%) than non-stroke victims (0.5%). However, after adjusting for confounding factors, they subsequently failed to detect a significant association in their multivariate analysis. In the more recent study, Domingos and Serra  used the IV Portuguese National Health Survey database to explore a possible association between nephrolithiasis and other major medical diseases, including cardiovascular diseases, diabetes, and obesity. Although after adjusting for age and body mass index, they found kidney stone formers to have a higher prevalence of hypertension, diabetes mellitus, myocardial infarction, and stroke (OR 1.330; 95% CI 1.015–1.743; P < 0.05) than non-stone formers, after further adjusting for hypertension and diabetes, the association between nephrolithiasis and stroke did not reach the significant level.
Therefore, although previous studies investigated the association between nephrolithiasis and stroke, the present observations between nephrolithiasis, ureterolithiasis, and stroke stand in contrast to their findings which failed to detect an association at the significant level. The present study also contributes to the body of knowledge surrounding the relationship between UC and stroke by further analysing the risk of stroke in patients with ureterolithiasis [12,13]. We propose that the associations observed in this study between UC and stroke reflect the underlying pathology of UC.
Nephrolithiasis has two aetiological characterisations, the primary being metabolic and the secondary including anatomical, infectious, medication- and disease-related causes. Previous studies have established the association between nephrolithiasis and metabolic syndrome, which is characterised by a cluster of features including dyslipidaemia, hyperglycaemia, hypertension, obesity, and insulin resistance . Such symptoms have been shown in a rat model of metabolic syndrome to precipitate changes in urinary constituents and to lead to an increased risk of uric acid and calcium stone formation . One prospective longitudinal study in particular has shown that uric acid stone formers have a significantly higher prevalence of diabetes and glucose intolerance, and an elevated serum triglyceride concentration compared with normal controls . As the results of the present study were detected after taking hyperlipidaemia, hypertension, diabetes, coronary heart disease, atrial fibrillation, heart failure, gout, and obesity into consideration, this study adds credence to the understanding of UC as a symptom of an underlying systemic metabolic disorder [11,19].
The understanding of nephrolithiasis and ureterolithiasis as different manifestations of the same underlying systemic metabolic disorder also helps to explain the lack of a statistical difference between the risks for stroke by stone location, and to promote the conception of UC as an indicator for systemic disorders characterised by adverse metabolic environments [19,22]. These same metabolic disturbances, which lead to the development of stone formation, may also contribute to stroke by proceeding through artery calcification .
The most common metabolic abnormalities present in stone formers include hypercalciuria, hyperoxaluria, and hypocitraturia . These abnormalities may proceed through the lack of effective calcification inhibitors, which is a factor in 80% of stone formers . This has been shown to be a common mechanism linking coronary artery calcification to calcium kidney stones . This mechanism has also been suggested to lead to intracranial artery calcification and predispose to stroke .
The present study's strengths include the use of a population-based dataset, which enabled the tracing of all the cases of UC and stroke during the study period. The large sample size afforded a considerable statistical advantage in detecting real differences between the two cohorts. However, the present study also had several limitations. Firstly, the diagnoses of UC and stroke relied on administrative claims data reported by physicians and hospitals. These may be less accurate than diagnoses made according to standardised criteria.
Secondly, some patient information on factors that may contribute to stroke, e.g. tobacco use, alcohol consumption, dietary habits, and body mass index, was not available through the administrative dataset. We also lacked information concerning the chewing of betel quid, a South East Asian peculiarity, which has been reported to produce a high incidence of hypercalciuria and hypocitraturia and thus may contribute to both UC and stroke . Therefore, the association between UC and stroke may be partially explained by the residual confounding by these factors.
Third, the present study did not analyse the effect that medication may have had on the development of stroke. Some medications, e.g. aspirin, may affect the development of stroke. Therefore, if more patients with UC took aspirin for their pain than patients with no UC, it is possible that on account of the protective effect of aspirin on stroke, the effect of UC on stroke might have been underestimated. But, in Taiwan this should not have been the case as NSAIDS and acetaminophen are used to treat pain in patients with UC. Furthermore, in Taiwan only patients with heart disease and those with prior stroke are advised to take aspirin prophylactically. But, this investigation adjusted for heart disease and excluded patients with prior stroke, and therefore long-term aspirin users would not have affected the results of the present study.
Fourth, the present study may have been partially victim to a surveillance bias, as patients with UC are more likely to have frequent outpatient clinic visits, which purely on account of increased exposure to the medical system may lead to an early detection of stroke or more frequent detections of silent stroke. This would also contribute to an ascertainment bias, in which patients with no UC may have been less likely to be included as stroke cases than patients with UC. Furthermore, as the NHI programme has a 98.4% coverage rate for the residents of Taiwan, 1.4% of Taiwan's residents are not covered and would thus not be included in the present study. As the intended population of this population-based study includes all the residents of Taiwan, the lack of 100% coverage results in a small ascertainment bias, as some members of the intended population (1.4%) are less likely to be included than others.
Finally, the risk for stroke may vary according to stone burden and composition; however, this information was not available for use in the present study.
In conclusion, after adjusting for patient demographic characteristics and co-morbid medical disorders, the present study detected an increased risk of stroke during the first 5 years after diagnosis of nephrolithiasis and/or ureterolithiasis. As stroke is a major cause of mortality and morbidity worldwide, and considering the relatively high and increasing incidence rates and low-mortality of UC, it is the hope of the authors that the existing preventative measures, e.g. blood pressure control and the use of statins and aspirin, be considered as strategies for risk reduction among the many UC survivors.
This study is based in part on data from the National Health Insurance Research Database provided by the Bureau of National Health Insurance, Department of Health, Taiwan and managed by the National Health Research Institutes. The Interpretations and conclusions contained herein do not represent those of the Bureau of National Health Insurance, Department of Health, or the National Health Research Institutes.
CONFLICT OF INTEREST
None of the authors have any financial interests to disclose.