Impact of Hypoglycemia on Morbidity, Mortality, and Resource Utilization in Gastrointestinal Stromal Tumor: A Nationwide Analysis

Abstract

Background: Non-islet cell tumor hypoglycemia is increasingly reported with gastrointestinal stromal tumors (GIST), but population-level estimates of its clinical impact are limited. We evaluated associations between hypoglycemia and inpatient outcomes among GIST hospitalizations. Methods: We conducted a retrospective cross-sectional study of the National Inpatient Sample (NIS) 2018–2020. Adult GIST discharges were identified by ICD-10-CM codes and stratified by hypoglycemia. Primary outcomes were in-hospital mortality and resource utilization—length of stay (LOS) and total hospital charge. Secondary outcomes included malnutrition, sepsis, ascites, peritonitis, bowel perforation, intestinal obstruction, gastrointestinal bleeding, and iron deficiency anemia. Analyses used survey-weighted logistic regression for binary outcomes and generalized linear models for continuous outcomes. A propensity score-matched sensitivity analysis balanced sepsis and malnutrition. Results: Among 61,725 GIST hospitalizations, 0.72% had hypoglycemia. Mortality was 12.6% with hypoglycemia vs. 3.1% without; adjusted odds of death were higher (aOR 4.16, 95% CI 2.06–8.37; p < 0.001). Hypoglycemia was also associated with malnutrition (aOR 5.63, 3.37–9.40), sepsis (aOR 4.00, 2.24–7.14), ascites (aOR 3.43, 1.63–7.19), and peritonitis (aOR 2.91, 1.17–7.22). LOS was 4.61 days longer on average (not significant; p = 0.185), and total hospital charge was $5218 higher (β = 19,116.8; p = 0.95). In the matched cohort, the mortality association attenuated but persisted (aOR 1.38, 1.27–1.49; p < 0.001); peritonitis remained significant (aOR 1.10, 1.04–1.17), intestinal obstruction (aOR 4.91, 3.44–7.05) and iron deficiency anemia (aOR 3.54, 1.62–7.74) became significant, while ascites and gastrointestinal bleeding were not significant. Conclusions: Hypoglycemia in GIST, although uncommon, marks a higher-risk inpatient trajectory with increased mortality and several complications; these signals largely persist after balancing severity proxies. Resource-use differences were directionally higher but not statistically significant. Recognition of hypoglycemia may aid risk stratification and inpatient management in GIST.

1. Introduction

Gastrointestinal stromal tumors (GISTs) are the most common mesenchymal neoplasms of the gastrointestinal tract. These tumors originate from the interstitial cells of Cajal, which regulate gut motility and function as the pacemakers of the gastrointestinal system [1,2]. GISTs are most frequently located in the stomach, followed by the small intestine, and less commonly in the colon and rectum [3]. The average annual incidence of GISTs is estimated to range from 5 to 15 cases per one million people [4].

Many GISTs are asymptomatic and discovered incidentally. When symptomatic, presentations vary based on tumor size and location. Common symptoms include abdominal pain, gastrointestinal bleeding, and nonspecific complaints such as early satiety or bloating [3,5]. In some patients, complications like obstruction, mucosal ulceration, or even intraperitoneal hemorrhage from tumor rupture may occur. However, in many cases, GISTs are small, often less than 2 cm, and remain unnoticed until imaging or endoscopic evaluation is performed for unrelated reasons [6].

The often vague and non-specific nature of GIST symptoms such as unexplained anemia or intermittent discomfort may lead to delayed diagnosis. These subtle presentations can also obscure paraneoplastic manifestations like hypoglycemia, which may go unrecognized until severe or refractory. While GIST is primarily known for its gastrointestinal manifestations, a growing number of case reports have associated it with paraneoplastic hypoglycemia, particularly non-islet cell tumor hypoglycemia (NICTH) [4,7,8,9]. For example, a recent case report described a patient with persistent hypoglycemia that remained unresponsive to intravenous dextrose infusion. Eventually, it was attributed to tumor-derived insulin-like growth factor-2 (IGF-2), particularly the high-molecular-weight form known as “big IGF-2, and was only resolved with glucocorticoids and tumor resection [10].

NICTH is a rare paraneoplastic syndrome with an estimated annual incidence of one per million people [11]. It occurs due to tumor secretion of IGF-2, which mimics insulin action and enhances peripheral glucose uptake while suppressing hepatic glucose output [7,11,12,13]. Though rare, its association with mesenchymal and epithelial tumors, including hepatocellular carcinoma and GIST, is increasingly recognized [3]. As with other causes of hypoglycemia in adults without diabetes, NICTH is a diagnosis of exclusion, established after documenting Whipple’s triad, excluding exogenous insulin/secretagogues and critical illnesses, and identifying a biochemical pattern of low insulin, low C-peptide, suppressed β-hydroxybutyrate, and either an elevated IGF-2:IGF-1 ratio or “big IGF-2” [14,15,16].

The mechanisms by which hypoglycemia occurs in GIST patients are multifactorial. It is speculated that IGF-2 production by the tumor suppresses pituitary growth hormone secretion and leads to an overall metabolic imbalance [17]. These patients often demonstrate low insulin, C-peptide, and β-hydroxybutyrate levels during hypoglycemia, supporting a diagnosis of NICTH. Some may experience recurrent or treatment-resistant episodes, which can be clinically challenging. Furthermore, as the tumor progresses or metastasizes, particularly to the liver, these metabolic disturbances may become more pronounced.

Although multiple case reports have described hypoglycemia in the context of GIST, large-scale data on its broader clinical significance remain limited [2,3,4,7,8,13,18,19,20,21,22,23,24]. Most prior reports consist of isolated case studies, which, though clinically valuable, do not reflect population-wide trends. These accounts typically describe rare or severe presentations and may not capture the full spectrum of disease encountered in routine inpatient care.

Evidence from studies in other malignancies, such as hepatocellular carcinoma, has shown that paraneoplastic hypoglycemia is associated with worse clinical outcomes, including higher mortality and complication rates [25]. Moreover, hypoglycemia in hospitalized patients has been consistently linked to prolonged hospital stays and increased healthcare charges, even after adjusting for comorbidities [26]. These observations raise the question of whether hypoglycemia in GIST may reflect a more aggressive tumor phenotype or serve as an early marker of disease severity.

To address this knowledge gap, we utilized the National Inpatient Sample (NIS) to evaluate whether hypoglycemia in patients hospitalized with GIST was associated with increased morbidity, mortality, and healthcare resource utilization [1].

2. Materials and Methods

2.1. Database

The NIS stands as the United States’ most extensive publicly accessible database for inpatient healthcare, encompassing data from various payers. It serves as a crucial tool for extensive data analysis, offering regional and national insights into inpatient usage, accessibility, charges, quality of care, insurance, demographic information, and clinical outcomes. Created through a collaborative effort between federal, state, and industry partners under the Agency for Healthcare Research and Quality’s Healthcare Cost and Utilization Project (HCUP), it includes information from approximately 20% of hospital admissions every year. The NIS employs a stratified sampling technique to ensure national representativeness and includes patient-level and hospital-level variables. Discharge weights provided within the database were applied to produce national estimates.

2.2. Study Population

The National Inpatient Sample (NIS) was queried for the years 2018–2020. Adult hospitalizations (age > 18 years) with a primary or secondary diagnosis of gastrointestinal stromal tumor (GIST) were identified using International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) codes (listed in Appendix A, Figure A1). This cohort was stratified based on the presence or absence of hypoglycemia.

Patients were included if they had a valid diagnosis of GIST with complete demographic and outcome data available. Exclusion criteria consisted of hospitalizations with missing demographic characteristics (e.g., age, sex, race, payer, income quartile) or incomplete outcome variables. To minimize misclassification, admissions without sufficient diagnostic coding or with invalid identifiers were also excluded. Comorbidities were classified using the Elixhauser Comorbidity Software developed by HCUP. The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of The Brooklyn Hospital (IRB ID -2340569-1 and date of approval 9 July 2025).

2.3. Study Outcomes

The primary outcomes studied were mortality and resource utilization, including LOS and total hospital charge. Secondary morbidity outcomes included iron deficiency anemia, malnutrition, sepsis, ascites, bowel perforation, peritonitis, intestinal obstruction, and gastrointestinal bleeding. The relevant ICD-10-CM codes used for identification are listed in the Appendix A and Appendix B below.

2.4. Statistical Analysis

Statistical analysis was performed using STATA/MP 17.0 (StataCorp, College Station, TX, USA). GIST hospitalizations were analyzed for outcomes including mortality, LOS, total hospital charge, and morbidity outcomes such as iron deficiency anemia, malnutrition, sepsis, ascites, bowel perforation, peritonitis, intestinal obstruction, and gastrointestinal bleeding. Categorical variables were compared using the chi-square test, while continuous variables were compared using the t-test. All analyses accounted for the complex survey design of the NIS using appropriate weighting, clustering, and stratification techniques to ensure valid national estimates.

Multivariate logistic regression was conducted to assess the association between hypoglycemia and outcomes in GIST hospitalizations. Confounding variables were selected based on clinical relevance and included sociodemographic characteristics, hospital-level factors, and comorbidities such as obesity, cannabis use, smoking, alcohol use, dyslipidemia, hypertension, chronic kidney disease, chronic liver disease, congestive heart failure, coronary artery disease, adrenal insufficiency, and insulinoma.

Adjusted odds ratios (aORs) with 95% confidence intervals (CIs) were reported for outcomes after multivariate analysis. For continuous outcomes such as length of stay (LOS) and hospital charge, generalized linear models with appropriate link functions were used. A p-value < 0.05 was considered statistically significant.

2.5. Propensity Score-Matched Sensitivity Analysis

To further address potential confounding, particularly from malnutrition and sepsis, which may both precipitate and reflect hypoglycemia, we performed a propensity score-matched (PSM) sensitivity analysis. The propensity score (hypoglycemia vs. no hypoglycemia) was estimated via logistic regression including malnutrition and sepsis in addition to the sociodemographic, hospital, and comorbidity covariates used in the primary models. Hospitalizations were matched on the propensity score; covariate balance was assessed using standardized mean differences (Appendix B, Figure A2). Outcomes in the matched cohort were re-estimated using survey-weighted logistic regression to obtain adjusted odds ratios (aORs) with 95% confidence intervals (CIs). This analysis was pre-specified as a robustness check to evaluate whether primary associations persisted after balancing on these high-impact clinical confounders.

As a sensitivity analysis, we performed a propensity score-matched comparison balancing sepsis and malnutrition. For binary outcomes, we also calculated E-values for the point estimate and the CI limit closest to the null to assess robustness to unmeasured confounding.

3. Results

3.1. Patient- and Hospital-Level Characteristics

Our study identified 61,725 hospitalizations with a diagnosis of GIST between 2018 and 2020. Of these, 0.72% had a concurrent diagnosis of hypoglycemia. Patients in the hypoglycemia cohort are noted to be younger, with a median age of 63 years compared to 66.6 years in those without hypoglycemia.

In terms of racial distribution, most GIST hospitalizations with hypoglycemia were noted in White patients (57.38%), followed by Black (26.23%), Hispanic (9.84%), and other racial groups (6.56%). This distribution was broadly similar to the non-hypoglycemia cohort, indicating no significant racial disparity in hypoglycemia prevalence among hospitalized GIST patients. The sex distribution was also comparable across cohorts, with females slightly more represented in the hypoglycemia group (50.82%) than in those without hypoglycemia (49.59%).

When evaluating socioeconomic status using ZIP code-based median household income quartiles, hypoglycemia prevalence was relatively consistent across all income groups. Interestingly, there was a slightly higher proportion of hypoglycemic GIST hospitalizations in the middle-income quartiles (27.84% each in the 26th–50th and 51st–75th percentiles).

Geographically, a greater proportion of hospitalizations with hypoglycemia occurred in the Southern United States (50.82%), which was also the most represented region in the non-hypoglycemia cohort (36.72%). Additionally, the majority of hospitalizations in both cohorts occurred in urban teaching hospitals (90.16% vs. 83.64%), likely reflecting referral patterns for complex oncologic care. Hospital bed-size and payer status were also comparable between groups.

Although none of the differences in demographic or hospital-level characteristics reached statistical significance, the observed trends may point toward underlying clinical or geographic patterns that deserve further exploration in larger prospective cohorts (

Table 1. Socio-demographics and hospital characteristics of GIST hospitalizations with and without Hypoglycemia.

Baseline Characteristics		GIST Hospitalizations with Hypoglycemia (%)	GIST Hospitalizations Without Hypoglycemia (%)	p Value
Age (in years)		63	66.63	--
Sex	Male	49.18	50.82	0.85
	Female	50.82	49.59
Race	White	57.38	59.9	0.71
	Black	26.23	21.7
	Hispanic	9.84	6.39
	Others	6.56	9.87
Quartile of median household income for zip code	0−25th	19.67	26.23	0.59
	26th−50th	27.87	23.62
	51st−75th	27.87	24.01
	76th−100th	24.59	26.14
Primary payer	Medicare	69.93	57.44	0.32
	Medicaid	14.75	9.02
	Private	19.67	29.03
	Others	1.64	4.5
Hospital teaching status and location	Rural	0.0	38.9	0.22
	Urban non-teaching	9.84	12.47
	Urban teaching	90.16	83.64
Hospital bed-size	Small	18.03	14.71	0.52
	Medium	29.51	25.81
	Large	52.46	59.48
Hospital region	North-east	14.75	20.49	0.13
	Mid-west	14.75	21.0
	South	50.82	36.72
	West	19.67	21.79

).

3.2. In-Hospital Complications

The prevalence of complications among GIST hospitalizations was compared between patients with and without hypoglycemia using chi-square analysis. Patients in the hypoglycemia group experienced a significantly higher in-hospital mortality rate, recorded at 12.6%, compared to 3.1% among those without hypoglycemia. After adjusting for potential confounders, hypoglycemia remained independently associated with a more than fourfold increase in the odds of in-hospital mortality (adjusted odds ratio [aOR]: 4.16; 95% CI: 2.06–8.37; p < 0.001) (

Table 2. Multivariate logistic regression analysis showing odds ratio of mortality outcomes of GIST hospitalizations stratified by hypoglycemia.

Complication	GIST with Hypoglycemia (%)	GIST Without Hypoglycemia (%)	Adjusted Odds Ratio (OR) *	95% Confidence Interval	p Value
Mortality	12.6	3.1	4.16	2.06–8.37	0.001

* OR < 1 means hospitalizations with hypoglycemia had lower odds of the outcome. Statistically significant protective associations indicated by p-value < 0.05. * Adjusted for socio-demographics, hospital characteristics, and other cardiac risk factors: obesity, cannabis, smoking, alcohol use, dyslipidemia, hypertension, chronic kidney and liver disorders, congestive heart failure, coronary artery disease, adrenal insufficiency, and insulinoma.

). Additionally, hypoglycemic episodes contributed to significantly longer hospitalizations, with an average increase in length of stay by 4.61 days. These findings suggest a considerable increase in clinical burden and resource utilization in patients presenting with both GIST and hypoglycemia.

In terms of secondary outcomes, multiple complications were significantly more common in the hypoglycemia group. Malnutrition was particularly prominent, present in 78.0% of patients with hypoglycemia, compared to only 14.2% in those without. This difference remained robust after multivariable adjustment (aOR: 5.63; 95% CI: 3.37–9.40; p < 0.001), indicating a strong independent association. Sepsis was also markedly elevated in the hypoglycemic cohort (30.8% vs. 7.6%), with aOR of 4.00 (95% CI: 2.24–7.14; p < 0.001). The risk of developing ascites (15.5% vs. 4.2%) and peritonitis (8.5% vs. 2.6%) was also significantly higher among those with hypoglycemia. Adjusted analyses showed aORs of 3.43 (95% CI: 1.63–7.19; p = 0.001) and 2.91 (95% CI: 1.17–7.22; p = 0.021), respectively, confirming that these complications were not merely due to baseline differences.

While trends toward higher proportions of bowel perforation, gastrointestinal bleeding, intestinal obstruction, and iron deficiency anemia were observed in the hypoglycemia group, these differences did not reach statistical significance following adjustment for confounders. For example, gastrointestinal bleeding occurred in 20.2% of hypoglycemic patients versus 12.1% of non-hypoglycemic patients, yet the adjusted odds ratio was 1.50 (95% CI: 0.76–2.96; p = 0.239), suggesting no statistically significant difference. Similarly, intestinal obstruction (4.7% vs. 3.0%) and iron deficiency anemia (17.1% vs. 12.3%) had elevated unadjusted rates in the hypoglycemia group but were not independently associated after multivariate analysis (p > 0.05 for both) (

Table 3. Comparison of influence of outcomes in hospitalized patients with gastrointestinal stromal tumor with and without hypoglycemia.

Complications	GIST with Hypoglycemia (%)	GIST Without Hypoglycemia (%)	Adjusted Odds Ratio (OR *)	95% Confidence Interval
Malnutrition	78	14.2	5.63	3.37–9.40
Sepsis	30.8	7.6	4.00	2.24–7.14
Ascites	15.5	4.2	3.43	1.63–7.19
Bowel Perforation	2.3	1.0	2.27	0.54–9.49
Peritonitis	8.5	2.6	2.91	1.17–7.22
Intestinal Obstruction	4.7	3.0	1.50	0.53–4.26
Gastrointestinal Bleed	20.2	12.1	1.50	0.76–2.96
Iron deficiency anemia	17.1	12.3	1.12	0.51–2.49

* OR > 1 means patients with hypoglycemia had higher odds of the outcome. Statistically significant protective associations indicated by p-value < 0.05. Statistically significant outcomes are indicated by bolded results.

, Figure 1).

In the propensity-matched cohort balancing malnutrition and sepsis, hypoglycemia remained associated with higher odds of mortality (aOR 1.38, 95% CI 1.27–1.49, p < 0.001), peritonitis (aOR 1.10, 95% CI 1.04–1.17, p < 0.001), intestinal obstruction (aOR 4.91, 95% CI 3.44–7.05, p < 0.001), and iron deficiency anemia (aOR 3.54, 95% CI 1.62–7.74, p < 0.001), whereas ascites and gastrointestinal bleeding were not statistically significant. These findings corroborate the direction of the primary models but with attenuation of the mortality effect, consistent with partial confounding by severity (

Table 4. Propensity score-matched outcomes among GIST hospitalizations with and without hypoglycemia (balanced on malnutrition and sepsis in addition to primary confounders).

Complications	Adjusted Odds Ratio of GIST Hospitalizations with and Without Hypoglycemia (aOR) *	95% Confidence Interval	p Value
Mortality	1.38	1.27–1.49	<0.001
Ascites	1.49	1.34–1.65	0.721
Peritonitis	1.10	1.04–1.17	<0.001
Intestinal Obstruction	4.91	3.44–7.05	<0.001
Gastrointestinal Bleed	0.68	0.23–1.93	0.470
Iron deficiency anemia	3.54	1.62–7.74	<0.001

* aORs from logistic regression in the matched cohort. Interpretation: OR > 1 indicates higher odds in hospitalizations with hypoglycemia. Two-sided p < 0.05 considered statistically significant.

).

4. Discussion

To our knowledge, this is the first national inpatient database analysis of the effects of hypoglycemia on GIST hospitalizations. The association of hypoglycemia with GIST hospitalizations is not uncommon. This relationship has been previously reported in the literature, mostly in case reports, where hypoglycemia is associated with GIST and presents either as the presenting complaint or occasionally during the treatment course [2,3,4,7,8,13,18,19,20,21,22,23,24].

Although regional differences did not reach statistical significance, we observed that a greater proportion of hypoglycemia-associated hospitalizations occurred in the Southern United States. This distribution may partly reflect the NIS sampling framework, which is designed to support regional as well as national estimates, but it could also align with population health patterns such as the higher burden of diabetes and related metabolic risk in parts of the South. Differences in healthcare access and referrals to urban teaching hospitals that manage complex oncologic care may additionally contribute. These observations remain exploratory and motivate future analyses that incorporate outpatient data and laboratory phenotyping to test potential geographic influences on vulnerability to hypoglycemia in GIST [27].

In this study, we sought to determine whether hypoglycemia had any significant correlation with tumor burden, as it may be associated with the production of IGF-2, which disrupts glucose balance and causes hypoglycemia [7,13]. Among hospitalizations for GIST, the presence of hypoglycemia was found to be significantly associated with increased odds of mortality after adjusting for possible confounders. These results were similar when compared with the study by Chida et al., a single-center observational study conducted over 12 years in Tokyo, which showed that recurrent hypoglycemic episodes were associated with poorer survival rates among GIST hospitalizations [28]. This consistency between our findings and those from earlier studies adds external validity and suggests that the association is not limited to specific populations or healthcare systems.

This study suggests that hypoglycemia could potentially serve as a prognostic marker in GIST hospitalizations. Several hypotheses have been proposed to explain hypoglycemia in GIST, including the paraneoplastic syndrome NICTH. When NICTH is associated with GIST, the tumor produces an altered form of IGF-2, known as “big IGF-2”, which is larger than normal IGF-2 [6,29]. Unlike the normal form, “big IGF-2” cannot form a complex with IGF-binding protein-3 (IGFBP-3), disrupting glucose regulation. As a result, less glucose is released by the liver, and more glucose is consumed by muscles, leading to hypoglycemia [30]. This could potentially correlate with tumor burden in GIST patients and may be one of the reasons for increased mortality among GIST hospitalizations. Furthermore, the association of hypoglycemia with higher mortality may reflect not only the biochemical effects of IGF-2 but also advanced or unresectable disease stages in which such paraneoplastic syndromes are more likely to manifest.

In this context, hypoglycemia may also serve as a surrogate indicator of advanced tumor biology. Prior studies have shown that paraneoplastic hypoglycemia is more frequently observed in patients with large tumor volumes or extensive hepatic metastases, both clinical indicators of late-stage or aggressive GIST. These phenotypes are often linked with higher levels of incompletely processed IGF-2, leading to systemic metabolic disturbances [31]. Therefore, hypoglycemia may reflect both disease burden and biological aggressiveness, offering potential utility as a clinical marker in stratifying risk among hospitalized GIST patients. Additionally, hypoglycemia may complicate the clinical course by limiting the use of specific systemic therapies or supportive interventions, indirectly influencing outcomes.

We also studied the impact of hypoglycemia on morbidity outcomes related to GIST. Hypoglycemia was significantly associated with worse outcomes such as malnutrition, sepsis, ascites, and peritonitis. Beyond the immediate inpatient course, malnutrition in established cancer is associated with poorer peri-treatment outcomes and prolonged recovery, a pattern that plausibly extends hospitalization duration [32]. This aligns with our cohort’s markedly higher malnutrition prevalence in the hypoglycemia group (78.0% vs. 14.2%) and the longer LOS observed (+4.61 days). Beyond inpatient factors, background metabolic risk and nutrition shape oncologic trajectories. Emerging data link steatosis/fasting hyperglycemia to higher GI cancer risk and progression, and low adherence to Mediterranean-style diet to greater GI cancer burden; malnutrition is associated with peri-treatment morbidity and likely longer hospitalization. These contextual factors support our observation of directionally higher LOS and charges in hypoglycemia-associated GIST admissions [32,33,34].

Increased peripheral utilization of glucose and disturbed glucose homeostasis, in addition to NICTH among these conditions, could contribute to hypoglycemia, correlating with worsening outcomes [35,36,37]. These complications likely reflect an overall higher physiological burden, possibly signifying more aggressive or advanced disease in hypoglycemic patients. On the other hand, complications such as GI bleeding did not show any statistically significant differences in outcomes among GIST hospitalizations. Episodes of hypoglycemia also led to an increased LOS by 4.61 days in GIST hospitalizations, indicating greater complexity of care and prolonged recovery periods. Increased hospital stay and resource utilization further highlight the potential economic impact of unrecognized or poorly managed hypoglycemia in this setting.

These findings are consistent with the observations made in the Chida et al. study, which similarly reported that hypoglycemia, especially with recurrent episodes, was more frequent in unresectable or metastatic GIST cases and was associated with a poorer prognosis. That study also confirmed the presence of NICTH in many of those cases through IGF-2 expression [28]. This alignment with prior literature strengthens the biological plausibility of our findings and underscores the importance of metabolic complications in GIST.

One plausible explanation for the clustering of complications including malnutrition, sepsis, ascites, and peritonitis is that hypoglycemia triggers a systemic inflammatory cascade. Controlled clamp studies and inpatient observations show that even a single hypoglycemic episode can induce leukocyte activation, increase pro-inflammatory cytokines (e.g., IL-6, TNF-α), and generate oxidative stress, with effects that may persist for days [3,38]. This sustained inflammatory milieu, superimposed on metabolic imbalance, likely impairs host defenses, oxygen delivery, and tissue repair, predisposing patients to secondary infections, delayed wound healing, and catabolic decline. Such dysregulation may be especially detrimental in patients with underlying malignancy and already limited physiologic reserve. The bidirectional relationship between metabolic instability and immune suppression warrants further mechanistic study. In our propensity score-matched analysis that balanced malnutrition and sepsis, the association with ascites attenuated and was no longer statistically significant, consistent with severity-related confounding rather than a direct effect. However, some complications including peritonitis, intestinal obstruction and iron deficiency anemia reached statistical significance.

Given these findings, hypoglycemia in GIST hospitalizations should not be dismissed as incidental. Instead, it may signal underlying disease severity or paraneoplastic activity. Clinicians should maintain a high index of suspicion, especially in patients with unexplained fatigue, functional decline, or systemic complications. Early glucose monitoring, and when appropriate, IGF-2 testing, may help identify these cases sooner and allow for timely intervention. Proactive recognition of hypoglycemia may also improve inpatient triage and support decisions around the intensity of care required.

We acknowledge several limitations of our study. Our study, conducted retrospectively using the NIS coding database, carries several inherent limitations. The dataset does not provide granular clinical information such as detailed medication usage, laboratory values, imaging findings, or pathology reports, all of which could meaningfully influence or clarify the outcomes studied. As a result, it was not feasible to assess the severity of GIST or hypoglycemia, correlate findings with tumor size, mutational status, or treatment modalities, or determine tumor staging. Similarly, the absence of temporal data precluded evaluation of the sequence of clinical events, making it unclear whether hypoglycemia precedes complications as an early marker of disease activity or develops as a consequence of advanced disease severity. The lack of information on molecular and biochemical parameters also prevented us from confirming IGF-2 expression or directly identifying cases of NICTH, thereby limiting mechanistic insight into the underlying pathophysiology.

The NIS is a discharge-level dataset; unique patient identifiers are not available, and the same individual may contribute multiple hospitalizations within a year. As a result, readmissions and inter-hospital transfers can appear as separate records, potentially over-representing patients with complex or refractory disease and modestly inflating complication counts or resource-use estimates if repeat utilizers cluster within the hypoglycemia group. Because we cannot distinguish index from subsequent admissions or account for within-patient correlation, our findings should be interpreted as associations at the discharge level, not per-patient risks.

Another important limitation relates to the source population. Because the NIS database primarily includes information from hospitalized patients, it may not fully represent the entire spectrum of GIST cases, many of which are managed in outpatient settings. Hospitalizations typically capture patients with more severe or complicated clinical courses, hypoglycemia being one such complication, potentially skewing the data toward advanced disease. This overrepresentation of more complex presentations may challenge the generalizability of our findings to the broader GIST population, where milder or asymptomatic cases may not be reflected.

Furthermore, the retrospective and cross-sectional nature of the NIS restricts causal inference. Without longitudinal follow-up, we could not establish whether hypoglycemia acts as an independent predictor of morbidity and mortality or whether it simply reflects disease burden, comorbid conditions, or treatment-related complications. Although we adjusted for multiple covariates, residual confounding remains possible, and the inability to account for unmeasured variables further limits definitive interpretation. In addition, we can only report associations using NIS and are unable to determine causality.

Despite these limitations, the strength of our analysis lies in its use of a large, nationally representative dataset, which provides robust insight into the association between hypoglycemia and adverse outcomes in hospitalized GIST patients. The consistent association observed underscores the clinical relevance of early recognition and management of hypoglycemia, as timely intervention may mitigate downstream complications. Looking ahead, prospective studies are warranted to validate our findings and to clarify the causal and mechanistic links between hypoglycemia and GIST outcomes. Integrating serial metabolic assessments, IGF-2 measurements, genomic tumor profiling, and treatment response data with clinical outcomes could determine whether hypoglycemia serves as a biomarker of disease progression or therapeutic response.

5. Conclusions

Hypoglycemia in patients with gastrointestinal stromal tumors (GIST) appears to have significant prognostic implications for both mortality and morbidity outcomes. While rare, its association with tumor burden and the production of IGF-2 suggests that hypoglycemia could serve as a valuable marker for assessing disease severity and guiding treatment strategies. Our findings suggest that routine screening for hypoglycemia in GIST admissions may offer early prognostic cues and aid in identifying patients at risk for worse outcomes, especially in advanced or unresectable diseases. Further prospective studies are needed to validate their potential role in clinical practice.