Relationship between serum alkaline phosphatase and poor 3-month prognosis in acute ischemic stroke patients with preserved renal function: results from Xi’an Stroke Registry Study of China

Acute ischemic stroke (AIS) is a sudden neurologic dysfunction caused by focal brain ischemia, and is the leading cause of worldwide death as well as disability-adjusted life years in China [1, 2]. It is well known that the risks of AIS are various and mainly include atherosclerosis, hypertension, diabetes mellitus, hyperlipemia, smoking and hyperhomocysteinemia [3]. Clinical studies have demonstrated in recent years that serum alkaline phosphatase (ALP) might be one of the independent risk factors causing ischemic stroke, since high serum ALP levels would promote the vascular calcification through catalyzing the hydrolysis of organic pyrophosphate and subsequent atherosclerosis [4]. Furthermore, elevated ALP levels were also considered to be related with systemic inflammation, malnutrition and metabolic syndrome, which may result in worse clinical outcomes in patients with strokes [4].

Previous studies revealed that renal impairment is closely associated with adverse events of cerebrovascular diseases [5, 6]. Serum ALP levels could induce vascular calcification [7, 8], and mineral disorders are closely associated with cardiovascular and cerebrovascular disease-related mortality in patients with end-stage renal disease [7, 8]. However, few studies have so far examined the relationship between serum ALP levels and poor early-term prognosis in stroke patients without impaired renal function. Moreover, different countries (or even different regions of the same country) may present inconsistent results due to differences in economic levels, dietary habits and lifestyle. In order to provide a basis for ALP-based therapeutics in strokes, we investigated whether serum ALP level is associated with poor 3-month prognosis in AIS patients with preserved renal function in the Xi’an district of China in this study.

In this multicenter perspective cohort study, we found that an elevated serum ALP level was an independent risk factor for a poor 3-month prognosis in AIS patients with preserved renal function in the Xi’an district of China. Such relationship was not significantly affected by age, sex, drinking, hypertension and leukocyte count (stratified by 10 × 10⁹/L). We also found that the relationship between ALP levels and a 3-month poor prognosis appeared as a J-shaped curve and the optimal threshold value was 90 U/L.

Previous studies showed that ALP levels could serve as a biomarker for the diagnosis and prognosis of stroke. Moreover, elevated serum ALP levels appeared to be associated with stroke severity, increased vascular deaths and recurrent vascular events, and an increased incidence of early mortality [15‐17]. The effect of ALP on prognosis in cardiovascular diseases has recently attracted increasing attention, and ALP level is associated with vascular calcification. Vascular calcification often leads to increased vascular sclerosis and decreased vascular compliance being closely related to atherosclerosis [18, 19].

In addition, elevated serum ALP levels were associated with the risk of cardiovascular morbidity and mortality in patients with renal failure [7, 8, 20]. However, as most previous studies pooled together stroke patients with and without renal failure, their results could not rule out the effect of abnormal renal function, even with adjustments for confounding factors. Although some studies recently focused on this issue, few of them have explored the relationship between serum ALP levels and a poor 3-month prognosis in AIS patients with preserved renal function.

In this study, the main patient variables were comparable between patients lost and not lost for follow-up at the end of 3 months (Supplementary table 1). Overall, the data showed a random pattern regarding the patients who were lost for follow-up in this study. Thus, the results of this study were not markedly affected by the loss for follow-up. The analyzed population was a good representation of the patients enrolled in this study. On the basis of ensuring sample representativeness, we attempted to explore the relationship between serum ALP levels and a poor 3-month prognosis in AIS patients with preserved renal function. Our findings showed that 5% of the patients had a poor prognosis at the end of the 3-month follow-up period. The analysis of multivariate-adjusted logistic regression showed that for every 10-U/L increase in the ALP level, the risk of a poor 3-month prognosis increased by 4%. The patients in the Q5 group had a 1.21-fold risk of a poor 3-month prognosis in comparison with those in Q4 group. These results imply that an elevated serum ALP level is an independent risk factor for a poor 3-month prognosis in AIS patients with preserved renal function. Clinicians should be concerned about rapidly elevated serum ALP levels, as such an increase may be associated with an even higher risk of a poor 3-month prognosis.

The relationship and optimal range of serum ALP levels regarding a poor 3-month prognosis were not clearly elucidated in previous studies. To clarify this issue, the threshold effect and dose–response relationship between ALP levels and a poor 3-month prognosis were systematically analyzed in this large-scale cohort study, and the results confirmed a nonlinear relationship. After being adjusted for potential confounders, the risk of a poor 3-month prognosis was significantly increased in Q1, Q2, Q3 and Q5 (with the highest risk in Q5) when compared to that in Q4 being considered as the reference group. Logistic curve fitting was performed using the ALP level as a continuous variable, and it showed a J-shaped-curve relationship between ALP levels and a poor 3-month prognosis (Fig. 2). The optimal ALP range was observed in Q4, and the optimal threshold value was 90 U/L (Table 4). Thus, values above 90 U/L may lead to an increase in the risk of a poor 3-month prognosis in this population.

A previous long-term study reported that the risk of all-cause mortality was 1.78-fold higher in the highest quintile of ALP (≥ 97 U/L) than in the lowest quintile (< 57 U/L). The curve fitting results showed that the relationship between ALP levels and all-cause mortality was close to a positive correlation straight line [21]. The Chinese National Stroke Registry Study showed that patients with their ALP levels being > 98 U/L had a 36% higher risk of all-cause mortality in a 1-year follow-up period, compared to those with ALP levels < 59 U/L. The curve fitting results also showed a linear relationship [22].

In this study, the poor 3-month prognosis of stroke caused by increased ALP level was similar to the previous studies, but low ALP level also suggested that the increased risk of 3-month prognosis of stroke patients was inconsistent with those in the previous results. This inconsistency may be due to different populations and durations of follow-up in various studies. We conducted this study with 3-month follow-up and excluded patients with eGFR ≤ 60 mL/min/1.73 m². In addition, differences in the study setting may have led to the inconsistent results because the patients in the present study were from the Xi’an district, which corresponds to the lower valley of the Wei River in the Central Shaanxi Plain of northwestern China. Apart from the different clinical characteristics from our included population, some previous basic investigations showed that low ALP levels could increase the amount of inhibitory pyrophosphate resulting in abnormal metabolism of pyridoxine-5 '-phosphate (the main form of vitamin B6) and hypomineralization of the skeleton. The loss of mineralization could cause a variety of symptoms including softening of bone, bowing and spontaneous breakage of bones, rickets and tooth defects [23, 24]. However, whether the reduced ALP level could be associated with poor 3-month prognosis in AIS patient with preserved renal function needs to be further explored and verified by subsequent relevant basic experiments. In addition, unlike those previous studies that included covariates based on statistical significance or clinical experience, the covariates were included as potential confounders in our study if they changed the estimates of ALP levels by > 10% with respect to a poor 3-month prognosis or if they were significantly associated with a poor 3-month prognosis. This may also be one of the reasons why the results are inconsistent with the previous studies. Our results suggested that clinicians could evaluate the probability of a poor 3-month prognosis according to the serum ALP levels at admission to strengthen the awareness of standardized treatment and complications and to reduce the risk of a poor early-term prognosis in AIS patients with preserved renal function in this district.

Previous studies have shown that ALP levels are associated with many risk factors for stroke, including age, sex, drinking [25], hypertension [26] and inflammation [23]. To further analyze the possible effects of these factors, the subgroup analysis in this study showed that the relationship between ALP levels (Q1-Q5) and a poor 3-month prognosis was not significantly altered by age, sex, drinking, hypertension and leukocyte count (stratified by 10 × 10⁹/L). These results suggested a relatively stable relationship and they provided a theoretical basis for clinicians to institute appropriate preventive managements and therapeutics for these patients.

The mechanisms underlying the relationship between elevated serum ALP levels and a poor 3-month prognosis are not fully known. There are several probable explanations to clarify the relationship. First, serum ALP may affect the permeability of the blood–brain barrier. ALP may also play an important role in the maintenance and integrity of the blood–brain barrier and in the transport of proteins across the barrier. An elevated serum ALP level may cause dysfunction in the transport of these proteins, leading to the breakdown of the blood–brain barrier and death of neurons [27‐29]. Second, ALP may be associated with neuroinflammation. An elevated serum ALP level is highly characteristic of an inflammatory response in many diseases. Previous studies have shown that ALP plays a critical role in suppressing neuroinflammation by interfering with immune activation and responses [30]. Infections of the respiratory and urinary tracts are very common after stroke [31, 32]. These may be associated with infection-induced inflammation and also be involved in ALP upregulation in patients with stroke. Third, ALP can alter vascular homeostasis. Previous studies have speculated that serum ALP is associated with vascular calcification and microcirculation disorders, causing the hardening of vascular walls, myocardial ischemia and further cerebrovascular dysfunction and cerebral ischemia [33, 34]. In addition, our study showed that the relationship between serum ALP levels and a poor 3-month prognosis followed a threshold effect with an increased risk at ALP levels > 90 U/L, especially in Q5, indicating that a rapid elevation in ALP levels may have a markedly deleterious effect and result in a poor prognosis. This also implies that an abnormal serum ALP level is a potential treatment target to improve poor early-term prognosis. However, this notion needs to be verified in subsequent experiments.

Our study has some limitations. First, the four selected hospitals were not randomly chosen for this study. All the selected hospitals were tertiary grade A but may not represent the status quo of AIS patients with preserved renal function at other community hospitals. Second, this study mainly focused on the relationship between serum ALP levels at admission and a poor 3-month prognosis, and failed to timely collect the stress level of ALP at symptom onset and the ALP level after discharge. In addition, the timepoint of each patient was not recorded at measurement. These deficiencies may have an impact on the results. Third, the TOAST (Trial of Org 10,172 in acute stroke treatment) classification of stroke was not included in our study data, so the different types of stroke were not analyzed. Fourth, Inflammation-related indicators such as C-reactive proteins and hypersensitive C-reactive proteins were not included in our data, that may result in the failure to fully consider the inflammatory response of patients in the analysis. Finally, possible complications such as biliary tract disease and liver disease may also affect serum ALP levels. The lack of data for these indicators is another limitation of our study.

In summary, this study indicates that increased serum ALP levels are an independent risk factor for a poor 3-month prognosis in AIS patients with preserved renal function, and the relationship between these two parameters revealed a J-shaped curve with an optimal threshold value of 90 U/L. Thus, ALP values higher than 90 U/L are associated with an increased risk of a poor 3-month prognosis in AIS patients with preserved renal function. Our findings should help physicians to assess the risk for a poor 3-month prognosis among this population and to provide an appropriate ALP range to guide clinical practice.