Introduction
Hyperglycemia is a public health problem with significant implications for the outcomes of hospitalized patients. It has been identified as a poor prognostic factor in various medical conditions, including cardiovascular diseases, severe infections, and systemic inflammatory states
1
.
Regionally, Latin America has a high burden of metabolic diseases, increasing the prevalence of hyperglycemia in emergency departments
2
–
3
.
Hyperglycemia is not only associated with worse outcomes in severe infections but also with a higher risk of cardiovascular events. Elevated glucose levels have been shown to negatively impact coagulation and hemostasis, promoting a prothrombotic state that increases mortality in various diseases
4
.
In patients with decompensated heart failure, a "U-shaped" relationship has been identified between hyperglycemia and mortality, with both excessively high and low blood glucose levels increasing the risk of death and rehospitalization
5
.
Similarly, in patients with acute myocardial infarction, admission hyperglycemia has been established as an independent predictor of both short- and long-term mortality
6
.
Furthermore, population-based studies have shown that the relationship between hyperglycemia and cardiovascular mortality in patients with diabetes or prediabetes follows an "L-shaped" pattern, indicating that even slight increases in glucose may negatively impact survival
7
.
The management of hyperglycemia in emergency departments varies considerably and remains a challenge. The American Diabetes Association guidelines
8
recommend an individualized approach based on continuous glucose monitoring and insulin use when necessary, prioritizing insulin analogs to reduce the risk of hypoglycemia. However, studies have shown that hospitals often use sliding scale insulin inappropriately, which can increase the risk of hypoglycemia and metabolic dysregulation
9
.
In this context, it is important to consider the impact of hyperglycemia on vulnerable populations, such as older adults with diabetes, who have a high prevalence of frailty, which is associated with worse cognitive and emotional function, increasing vulnerability to adverse outcomes
10
.
Despite the existing evidence on the association between hyperglycemia and poor prognosis in various clinical conditions, there is still limited information in the Latin American context regarding emergency management strategies and their impact on mortality and complications.
Therefore, the aim of this study is to determine the management of hyperglycemia in the emergency department and its relationship with glycemic control, as well as the impact of glycemic control on mortality and adverse outcomes in three national hospitals in Peru during the period from August to December 2022.
Methods
Study design and setting
An observational, analytical, retrospective cohort study was conducted. The study setting included three national referral hospitals located in Metropolitan Lima: Hospital Nacional Hipólito Unanue, Hospital Nacional Dos de Mayo (both under the Ministry of Health), and Hospital Edgardo Rebagliati Martins, part of the social security system (EsSalud). Data collection covered the period from August to December 2022.
Population and sample
The study population included patients over 18 years of age who were admitted to the emergency internal medicine service for hyperglycemia at the mentioned hospitals. A non-probabilistic convenience sampling method was used. Inclusion criteria were: admission to the emergency department for hyperglycemia, serum glucose level over 180 mg/dL, with or without a history of diabetes mellitus (DM), and a minimum stay of 24 hours in the emergency department. Patients were excluded if they had incomplete clinical records or lacked information on the insulin regimen and clinical outcomes.
Variables and instruments
The dependent variables were: need for mechanical ventilation, acute kidney injury (AKI) requiring hemodialysis, death, and a composite outcome including the occurrence of at least one of these three events, defined after the first 24 hours of admission to avoid biases related to early clinical evolution or decisions. Prolonged hospital stay was analyzed separately and was not part of the composite outcome. Additionally, cases of decompensated chronic kidney disease (CKD) were excluded from the hemodialysis outcome.
The main independent variable was glycemic control at 24 hours after the initiation of insulin treatment, defined as a glucose level ≤180 mg/dL, according to international recommendations
8
, for hospitalized patients. Data collected included sociodemographic factors (age, sex, marital status, type of insulin regimen), medical history (diabetes mellitus, hypertension, neoplasms, diabetic foot, infections, cerebrovascular events), and reasons for admission associated with or not related to hyperglycemia (such as diabetic ketoacidosis, hyperosmolar or mixed states). Vital signs upon admission (blood pressure, heart rate, respiratory rate, temperature) and biomarkers (initial glucose and at 24 hours, hemoglobin, leukocytes, neutrophils, lymphocytes, platelets, C-reactive protein, and arterial gases) were also recorded.
Hypoxemia was defined by a SatO₂/FiO₂ ratio less than 315. Anemia was considered if hemoglobin was below 12 g/dL in females or 13 g/dL in males. Platelets were categorized as thrombocytopenia (<150×10³/μL), normal (150–400×10³/μL), and thrombocytosis (>450×10³/μL). Leukocytes were classified as leukopenia (<4×10³/μL), normal (4–11×10³/μL), and leukocytosis (>11×10³/μL); neutrophils as neutropenia (<1.5×10³/μL), normal (1.5–7.7×10³/μL), and neutrophilia (>7.7×10³/μL); and lymphocytes as lymphopenia (<1×10³/μL), normal (1–4.8×10³/μL), and lymphocytosis (>4.8×10³/μL). The acid-base status was defined based on pH, HCO₃, and pCO₂ as metabolic or respiratory acidosis or alkalosis, or normal status.
Procedures
Clinical and demographic data were extracted from both electronic and physical medical records of the patients, using a pre-prepared data collection form. The medical progress notes, medical orders, laboratory results, and hospitalization reports were reviewed to obtain complete and accurate information on each case.
Statistical analysis
The data were entered into a database in Microsoft Excel and then analyzed using the statistical software STATA version 16. For descriptive analysis, central tendency and dispersion measures were used for quantitative variables, and absolute and relative frequencies for categorical variables. In the bivariate analysis, the Student's T-test was applied for comparing means of numerical variables, and the Chi-square test for comparing proportions in categorical variables.
Multivariate analysis was conducted using a Poisson regression model with robust variances, including the variables that showed significant association in the bivariate analysis, as well as the primary independent variable of interest, which was the achievement of glycemic control at 24 hours. To avoid collinearity, variables with clinical overlap or shared origins from the same pathophysiological process (e.g., diabetic foot and diabetes) were excluded. For hematological parameters from the leukocyte count (total leukocytes, neutrophils, and lymphocytes), if both leukocytes and neutrophils or leukocytes and lymphocytes showed significant association with the outcome, neutrophils or lymphocytes were prioritized, excluding leukocytes due to their high correlation with these cell subtypes. A p-value of <0.05 was considered statistically significant, and the crude (RR) and adjusted (aRR) relative risks with their respective 95% confidence intervals were estimated.
Ethical considerations
The study was conducted in accordance with the ethical principles outlined in the Declaration of Helsinki and was approved by the ethics committees of the three participating hospitals. The confidentiality and anonymity of the patients were ensured, maintaining their privacy and the integrity of their personal data throughout the study.
RESULTS
Table 1 shows that the median age of the patients was 60.0 years (interquartile range (IQR): 50.0-70.0), and 52.8% were male. The majority had a history of diabetes mellitus (85.1%), and nearly one-third had hypertension (32.6%). Regarding outcomes, 56.6% had a prolonged hospital stay, and 41.0% required admission to critical care units. The median duration of hospital stay was 12.0 days (IQR: 8.0-17.0), and the mortality rate reached 4.0%.
Table 1. General characteristics of patients admitted for hyperglycemia in the emergency department in three Peruvian hospitals
Clinical and demographic characteristics |
Total (N=730) |
Age | 60.0 (50.0–70.0) |
Sex | |
Male | 385 (52.8%) |
Female | 344 (47.2%) |
Occupation | |
Housewife | 218 (29.9%) |
Merchant | 127 (17.4%) |
Retired | 90 (12.3%) |
None | 23 (3.2%) |
Others | 271 (37.1%) |
Marital status | |
Single | 221 (30.3%) |
Married | 399 (54.7%) |
Divorced | 41 (5.6%) |
Widowed | 69 (9.5%) |
Medical history | |
Diabetes mellitus | 621 (85.1%) |
Hypertension (HTN) | 238 (32.6%) |
Cirrhosis | 16 (2.2%) |
Chronic lung disease | 15 (2.1%) |
Neoplasms | 156 (21.4%) |
Other | 117 (16.0%) |
Adverse outcomes | |
Prolonged stay – No | 317 (43.4%) |
Prolonged stay – Yes | 413 (56.6%) |
Hospital stay (days) | 12.0 (8.0–17.0) |
Mortality – Alive | 701 (96.0%) |
Mortality – Deceased | 29 (4.0%) |
Admission to critical care units | |
No | 431 (59.0%) |
Yes | 299 (41.0%) |
Need for mechanical ventilation | |
No | 678 (92.9%) |
Yes | 52 (7.1%) |
AKI with need for hemodialysis | |
No | 694 (95.1%) |
Yes | 36 (4.9%) |
AKI: Acute kidney injury. HTN: Hypertension.
The most frequent reasons for admission are detailed in Table 2, where it is observed that hyperglycemic crises accounted for 30.4% of the cases, followed by diabetic foot (20.8%), urinary tract infection (20.5%), and skin and soft tissue infection (16.0%). Other relevant diagnoses included pneumonia (13.8%) and cerebrovascular disorder (8.1%), while less frequent causes were acute coronary syndrome (1.4%) and hypertension in the context of urgency or emergency (1.0%).
Reasons for admission of patients with hyperglycemia treated in the emergency department of three Peruvian hospitals
Reason for Admission |
Total, n (%) |
Hyperglycemic crises* | 222 (30,4) |
Diabetic foot | 152 (20,8) |
Urinary tract infection | 150 (20,5) |
Skin and soft tissue infection | 117 (16,0) |
Pneumonia | 101 (13,8) |
Cerebrovascular disorder | 59 (8,1) |
Others | 59 (8,1) |
Acute respiratory failure | 26 (3,6) |
Abdominal sepsis/acute gastroenteritis | 21 (2,9) |
Decompensated chronic kidney disease | 21 (2,9) |
Encephalopathy/sensory disturbance | 19 (2,6) |
Acute coronary syndrome / acute myocardial infarction | 10 (1,4) |
Abdominal pain syndrome | 9 (1,2) |
Hypertension (emergency/urgency) | 7 (1,0) |
Liver failure | 2 (0,3) |
* Hyperglycemic crises include diabetic ketoacidosis (n=182), hyperosmolar state (n=13), and mixed state (n=27).
Of the total patients who received a fixed-dose regimen (n = 253), 53.0% achieved glycemic control, while 47.0% did not achieve it. In contrast, among patients managed with a sliding-scale regimen (n = 477), only 41.1% achieved glycemic control within 24 hours, while 58.9% did not reach this goal. This difference was statistically significant (p=0.002).
In Table 3, factors associated with prolonged hospital stay were identified in the adjusted analysis, including the presence of neoplasms (aRR: 0.69; 95% CI: 0.56-0.85), hypoxemia (aRR: 1.29; 95% CI: 1.07-1.56), neutrophilia (aRR: 1.27; 95% CI: 1.08-1.49), metabolic alkalosis (aRR: 1.30; 95% CI: 1.09-1.55), respiratory acidosis (aRR: 1.35; 95% CI: 1.06-1.72), and C-reactive protein levels (aRR: 0.55; 95% CI: 0.46-0.67).
Table 3. Risk factors for prolonged hospital stay in patients with hyperglycemia treated in the emergency department of three Peruvian hospitals.
Variable |
n (%) |
p-value* |
RR (IC95%) |
p-value* |
RRa (IC95%) |
p-value* |
No prolonged stay (n=317) |
With prolonged stay (n=413) |
Age (median, IQR) | 60.0 (51.0–70.0) | 60.0 (50.0–69.0) | 0.910 | 1.00 (0.99–1.00) | 0.688 | - | - |
Sex |
Male | 167 (43.4%) | 218 (56.6%) | 0.950 | Ref. | Ref. | - | - |
Female | 150 (43.6%) | 194 (56.4%) | 1.00 (0.88–1.13) | 0.951 | - | - |
Marital status |
Single | 97 (43.9%) | 124 (56.1%) | 0.880 | Ref. | Ref. | - | - |
Married | 174 (43.6%) | 225 (56.4%) | 1.00 (0.87–1.16) | 0.946 | - | - |
Divorced | 19 (46.3%) | 22 (53.7%) | 0.96 (0.70–1.30) | 0.776 | - | - |
Widowed | 27 (39.1%) | 42 (60.9%) | 1.08 (0.87–1.36) | 0.473 | - | - |
Diabetes mellitus |
No | 40 (36.7%) | 69 (63.3%) | 0.120 | Ref. | Ref. | - | - |
Yes | 277 (44.6%) | 344 (55.4%) | 0.88 (0.75–1.03) | 0.101 | - | - |
Hypertension |
No | 220 (44.7%) | 272 (55.3%) | 0.310 | Ref. | Ref. | - | - |
Yes | 97 (40.8%) | 141 (59.2%) | 1.07 (0.94–1.22) | 0.305 | - | - |
Neoplasms |
No | 229 (39.9%) | 345 (60.1%) | <0.001 | Ref. | Ref. | - | - |
Yes | 88 (56.4%) | 68 (43.6%) | 0.73 (0.60–0.88) | 0.001 | 0.69 (0.56–0.85) | <0.001 |
Hyperglycemic crises |
No | 221 (43.5%) | 287 (56.5%) | 0.950 | Ref. | Ref. | - | - |
Yes | 96 (43.2%) | 126 (56.8%) | 1.00 (0.88–1.15) | 0.948 | - | - |
Diabetic foot |
No | 257 (44.5%) | 321 (55.5%) | 0.270 | Ref. | Ref. | - | - |
Yes | 60 (39.5%) | 92 (60.5%) | 1.09 (0.94–1.26) | 0.254 | - | - |
Urinary tract infection |
No | 251 (43.3%) | 329 (56.7%) | 0.870 | Ref. | Ref. | - | - |
Yes | 66 (44.0%) | 84 (56.0%) | 0.99 (0.84–1.16) | 0.874 | - | - |
Skin and soft tissue infection |
No |
264 (43,1%) |
349 (56,9%) |
0,660 |
Ref. |
Ref. |
- |
- |
Yes |
53 (45,3%) |
64 (54,7%) |
0,96 (0,80–1,15) |
0,661 |
- |
- |
Pneumonia |
No |
280 (44,5%) |
349 (55,5%) |
0,140 |
Ref. |
Ref. |
- |
- |
Yes |
37 (36,6%) |
64 (63,4%) |
1,14 (0,97–1,35) |
0,113 |
- |
- |
Cerebrovascular disorder |
No |
293 (43,7%) |
378 (56,3%) |
0,660 |
Ref. |
Ref. |
- |
- |
Yes |
24 (40,7%) |
35 (59,3%) |
1,05 (0,84–1,31) |
0,648 |
- |
- |
First glucose level |
294,0 (217,0–416,0) |
296,4 (226,1–409,5) |
0,440 |
1,00 (1,00–1,00) |
0,430 |
- |
- |
Delta glucose at 24 hours |
-112,5 (-226,0–-47,0) |
-127,2 (-234,0–-69,0) |
0,068 |
1,00 (1,00–1,00) |
0,200 |
- |
- |
Systolic BP |
120,0 (100,0–135,0) |
120,0 (100,0–130,0) |
0,720 |
1,00 (1,00–1,00) |
0,705 |
- |
- |
Diastolic BP |
70,0 (60,0–80,0) |
70,0 (60,0–80,0) |
0,330 |
1,00 (1,00–1,00) |
0,285 |
- |
- |
Heart rate |
89,0 (78,0–102,0) |
88,5 (78,0–102,0) |
0,930 |
1,00 (1,00–1,00) |
0,633 |
- |
- |
Respiratory rate |
20,0 (18,0–22,0) |
20,0 (18,0–22,0) |
0,930 |
1,00 (0,99–1,01) |
0,799 |
- |
- |
Temperature |
37,0 (36,2–37,0) |
37,0 (36,0–37,0) |
0,160 |
1,00 (0,97–1,03) |
0,794 |
- |
- |
Hypoxemia |
No |
290 (44,8%) |
357 (55,2%) |
0,002 |
Ref. |
Ref. |
- |
- |
Yes |
15 (24,2%) |
47 (75,8%) |
1,37 (1,17–1,61) |
<0,001 |
1,29 (1,07–1,56) |
<0,007 |
SaFi Index (SaO₂/FiO₂) |
415,2 (342,9–485,7) |
400,0 (325,8–495,2) |
0,280 |
1,00 (1,00–1,00) |
0,840 |
- |
- |
Anemia |
No |
148 (43,8%) |
190 (56,2%) |
0,850 |
Ref. |
Ref. |
- |
- |
Yes |
169 (43,1%) |
223 (56,9%) |
1,01 (0,89–1,15) |
0,855 |
- |
- |
Platelets |
Thrombocytopenia |
19 (42,2%) |
26 (57,8%) |
0,500 |
1,00 (0,77–1,29) |
0,977 |
- |
- |
Normal |
223 (42,0%) |
308 (58,0%) |
Ref. |
Ref. |
- |
- |
Thrombocytosis |
52 (48,1%) |
56 (51,9%) |
0,89 (0,74–1,09) |
0,262 |
- |
- |
Leukocytes |
Leukopenia |
2 (28,6%) |
5 (71,4%) |
0,190 |
1,35 (0,84–2,20) |
0,217 |
- |
- |
Normal |
137 (47,2%) |
153 (52,8%) |
0,74 (0,46–1,20) |
0,217 |
- |
- |
Leukocytosis |
178 (41,1%) |
255 (58,9%) |
1,12 (0,98–1,28) |
0,109 |
- |
- |
Neutrophils |
Neutropenia |
5 (41,7%) |
7 (58,3%) |
0,097 |
1,15 (0,70–1,88) |
0,586 |
1,35 (0,85–2,14) |
0,211 |
Normal |
116 (49,2%) |
120 (50,8%) |
Ref. |
Ref. |
- |
- |
Neutrophilia |
196 (40,7%) |
286 (59,3%) |
1,17 (1,01–1,35) |
0,038 |
1,27 (1,08–1,49) |
0,004 |
Lymphocytes |
Lymphopenia |
92 (50,8%) |
89 (49,2%) |
0,063 |
0,84 (0,71–0,99) |
0,033 |
0,91 (0,76–1,09) |
0,292 |
Normal |
213 (41,2%) |
304 (58,8%) |
Ref. |
Ref. |
- |
- |
Lymphocytosis |
12 (37,5%) |
20 (62,5%) |
1,06 (0,80–1,40) |
0,667 |
0,98 (0,71–1,36) |
0,900 |
Metabolic status according to ABG |
Metabolic acidosis |
76 (42,9%) |
101 (57,1%) |
0,004 |
1,15 (0,96–1,36) |
0,121 |
1,01 (0,84–1,21) |
0,945 |
Metabolic alkalosis |
48 (33,1%) |
97 (66,9%) |
1,34 (1,14–1,58) |
<0,001 |
1,30 (1,09–1,55) |
0,004 |
Respiratory acidosis |
8 (28,6%) |
20 (71,4%) |
1,43 (1,10–1,86) |
0,007 |
1,35 (1,06–1,72) |
0,016 |
Respiratory alkalosis |
33 (49,3%) |
34 (50,7%) |
1,02 (0,78–1,32) |
0,892 |
1,09 (0,83–1,43) |
0,520 |
Normal |
149 (50,2%) |
148 (49,8%) |
Ref. |
Ref. |
- |
- |
C-Reactive protein |
9,5 (1,3–47,3) |
10,3 (2,0–24,0) |
0,340 |
1,00 (1,00–1,00) |
0,001 |
0,55 (0,46–0,67) |
<0,001 |
* Chi-square test (categorical independent variable) or Mann-Whitney U test (numerical independent variable).
RR: Relative Risk. aRR: Adjusted Relative Risk. HTN: Hypertension. BP: Blood Pressure. ABG: Arterial Blood Gas analysis.
In Table 4, it can be observed that, in the adjusted analysis, the factors associated with a higher risk of presenting the combined outcome were the presence of neoplasms (aRR: 1.71; 95% CI: 1.10-2.67), pneumonia (aRR: 1.84; 95% CI: 1.14-2.97), cerebrovascular disorder (aRR: 1.63; 95% CI: 1.05-2.51), elevated temperature (aRR: 1.34; 95% CI: 1.07-1.66), hypoxemia (aRR: 2.66; 95% CI: 1.80-3.93), thrombocytopenia (aRR: 2.21; 95% CI: 1.32-3.69), neutrophilia (aRR: 1.65; 95% CI: 1.07-2.55), lymphopenia (aRR: 0.61; 95% CI: 0.39-0.94), metabolic acidosis (aRR: 1.87; 95% CI: 1.17-2.99), and respiratory acidosis (aRR: 3.12; 95% CI: 1.95-4.97).
Table 4. Risk factors for combined outcome in patients with hyperglycemia admitted to the emergency department of three Peruvian hospitals.
|
Without Combined Outcome (n=634) |
With Combined Outcome (n=96) |
p-value |
RR (IC95%) |
p-value |
aRR (IC95%) |
p-value |
Age |
60,0 (50,0–70,0) |
62,5 (51,0–70,5) |
0,380 |
1,00 (0,99–1,02) |
0,373 |
- |
- |
Sex |
Male |
336 (87,3%) |
49 (12,7%) |
0,710 |
Ref. |
Ref. |
- |
- |
Female |
297 (86,3%) |
47 (13,7%) |
1,07 (0,74–1,56) |
0,709 |
- |
- |
Marital Status |
Single |
196 (88,7%) |
25 (11,3%) |
0,510 |
Ref. |
Ref. |
- |
- |
Married |
344 (86,2%) |
55 (13,8%) |
1,22 (0,78–1,90) |
0,382 |
- |
- |
Divorces |
37 (90,2%) |
4 (9,8%) |
0,86 (0,32–2,35) |
0,772 |
- |
- |
Widowed/td>
| 57 (82,6%) |
12 (17,4%) |
1,54 (0,82–2,90) |
0,183 |
- |
- |
Diabetes mellitus |
No |
87 (79,8%) |
22 (20,2%) |
0,018 |
Ref. |
Ref. |
- |
- |
Yes |
547 (88,1%) |
74 (11,9%) |
0,59 (0,38–0,91) |
0,016 |
1,11 (0,77–1,59) |
0,583 |
HTN |
No |
422 (85,8%) |
70 (14,2%) |
0,220 |
Ref. |
Ref. |
- |
- |
Yes |
212 (89,1%) |
26 (10,9%) |
0,77 (0,50–1,17) |
0,211 |
- |
- |
Neoplasms |
No |
509 (88,7%) |
65 (11,3%) |
0,005 |
Ref. |
Ref. |
- |
- |
Yes |
125 (80,1%) |
31 (19,9%) |
1,75 (1,19–2,59) |
0,005 |
1,71 (1,10–2,67) |
0,017 |
Hyperglycemic crisis |
No |
434 (85,4%) |
74 (14,6%) |
0,087 |
Ref. |
Ref. |
- |
- |
Yes |
200 (90,1%) |
22 (9,9%) |
0,68 (0,43–1,07) |
0,093 |
- |
- |
Diabetic foot |
No |
492 (85,1%) |
86 (14,9%) |
0,007 |
Ref. |
Ref. |
- |
- |
Yes |
142 (93,4%) |
10 (6,6%) |
0,44 (0,24–0,83) |
0,011 |
- |
- |
Urinary tract infection |
No |
495 (85,3%) |
85 (14,7%) |
0,018 |
Ref. |
Ref. |
- |
- |
Yes |
139 (92,7%) |
11 (7,3%) |
0,50 (0,27–0,91) |
0,024 |
0,71 (0,41–1,22) |
0,215 |
Skin and soft tissue infection |
No |
521 (85,0%) |
92 (15,0%) |
<0,001 |
Ref. |
Ref. |
- |
- |
Yes |
113 (96,6%) |
4 (3,4%) |
0,23 (0,09–0,61) |
0,003 |
0,33 (0,11–1,03) |
0,056 |
Pneumonia |
No |
563 (89,5%) |
66 (10,5%) |
<0,001 |
Ref. |
Ref. |
- |
- |
Yes |
71 (70,3%) |
30 (29,7%) |
2,83 (1,94–4,12) |
<0,001 |
1,84 (1,14–2,97) |
0,013 |
Cerebrovascular disorder |
No |
590 (87,9%) |
81 (12,1%) |
0,004 |
Ref. |
Ref. |
- |
- |
Yes |
44 (74,6%) |
15 (25,4%) |
2,10 (1,30–3,41) |
0,002 |
1,63 (1,05–2,51) |
0,028 |
First glucose |
297,0 (223,0-416,0) |
273,0 (218,5-372,7) |
0,260 |
1,00 (1,00-1,00) |
0,668 |
- |
- |
Delta glucose at 24 hours |
-123,0 (-234,0--57,6) |
-112,0 (-216,7--67,6) |
0,900 |
1,00 (1,00-1,00) |
0,977 |
- |
- |
Systolic BP |
120,0 (100,0-134,0) |
110,0 (90,0-129,0) |
0,011 |
1,00 (0,99-1,00) |
0,328 |
- |
- |
Diastolic BP |
70,0 (60,0-80,0) |
65,0 (57,5-80,0) |
0,039 |
0,99 (0,98-1,00) |
0,322 |
- |
- |
Heart rate |
89,0 (78,0-102,0) |
89,0 (78,0-103,0) |
0,680 |
1,00 (0,99-1,01) |
0,738 |
- |
- |
Respiratory rate |
20,0 (18,0-22,0) |
20,0 (18,0-27,0) |
0,003 |
1,04 (1,01-1,06) |
0,004 |
1,00 (0,97-1,04) |
0,803 |
Temperature |
37,0 (36,0-37,0) |
37,0 (36,7-37,5) |
<0,001 |
1,42 (1,18-1,70) |
<0,001 |
1,34 (1,07-1,66) |
0,009 |
Hipoxemia |
|
<0,001 |
|
No |
585 (90,4%) |
62 (9,6%) |
|
Ref. |
Ref. |
- |
- |
Yes |
29 (46,8%) |
33 (53,2%) |
|
5,55 (3,98-7,75) |
<0,001 |
2,66 (1,80-3,93) |
<0,001 |
SaFi Index (SaO₂/FiO₂) |
414,3 (351,0-495,2) |
370,0 (202,8-466,7) |
<0,001 |
1,00 (1,00-1,00) |
0,492 |
|
|
Anemia |
|
0,100 |
|
No |
301 (89,1%) |
37 (10,9%) |
|
Ref. |
Ref. |
- |
- |
Yes |
333 (84,9%) |
59 (15,1%) |
|
1,37 (0,94-2,01) |
0,105 |
- |
- |
Platelets |
|
0,029 |
|
Thrombocytopenia |
33 (73,3%) |
12 (26,7%) |
|
2,02 (1,19-3,44) |
0,009 |
2,21 (1,32-3,69) |
0,002 |
Normal |
461 (86,8%) |
70 (13,2%) |
|
Ref. |
Ref. |
- |
- |
Thrombocytosis |
96 (88,9%) |
12 (11,1%) |
|
0,84 (0,47-1,50) |
0,561 |
1,11 (0,63-1,93) |
0,722 |
Leukocytes |
|
0,076 |
|
Leukopenia |
6 (85,7%) |
1 (14,3%) |
|
1,48 (0,23-9,41) |
0,678 |
- |
- |
Normal |
262 (90,3%) |
28 (9,7%) |
|
Ref. |
Ref. |
- |
- |
Leukocytosis |
366 (84,5%) |
67 (15,5%) |
|
1,08 (0,17-6,75) |
0,932 |
- |
- |
Neutrófilos |
|
0,049 |
|
Neutropenia |
11 (91,7%) |
1 (8,3%) |
|
0,94 (0,14-6,40) |
0,947 |
0,83 (0,18-3,89) |
0,809 |
Normal |
215 (91,1%) |
21 (8,9%) |
|
Ref. |
Ref. |
- |
- |
Neutrophilia |
408 (84,6%) |
74 (15,4%) |
|
1,73 (1,09-2,73) |
0,020 |
1,65 (1,07-2,55) |
0,023 |
Lymphocytes |
|
0,120 |
|
Lymphopenia |
157 (86,7%) |
24 (13,3%) |
|
1,07 (0,69-1,66) |
0,758 |
0,61 (0,39-0,94) |
0,026 |
Normal |
453 (87,6%) |
64 (12,4%) |
|
Ref. |
Ref. |
- |
- |
Lymphocytosis |
24 (75,0%) |
8 (25,0%) |
|
2,02 (1,06-3,84) |
0,032 |
0,83 (0,48-1,46) |
0,527 |
Metabolic status according to ABG |
|
<0,001 |
|
Metabolic acidosis |
144 (81,4%) |
33 (18,6%) |
|
1,85 (1,17-2,92) |
0,009 |
1,87 (1,17-2,99) |
0,009 |
Metabolic alkalosis |
136 (93,8%) |
9 (6,2%) |
|
0,61 (0,30-1,26) |
0,184 |
1,18 (0,56-2,49) |
0,656 |
Respiratory acidosis |
13 (46,4%) |
15 (53,6%) |
|
5,30 (3,27-8,61) |
<0,001 |
3,12 (1,95-4,97) |
<0,001 |
Respiratory alkalosis |
60 (89,6%) |
7 (10,4%) |
|
1,03 (0,47-2,25) |
0,932 |
0,67 (0,28-1,63) |
0,376 |
Normal |
267 (89,9%) |
30 (10,1%) |
|
Ref. |
Ref. |
- |
- |
C-Reactive protein |
10,5 (1,9-28,7) |
6,1 (1,0-21,1) |
0,053 |
1,00 (0,99-1,00) |
0,073 |
- |
- |
* Chi-square test (categorical independent variable) or Mann Whitney U test (numerical independent variable). RR: Relative risk. aRR: adjusted relative risk. HTN: hypertension. BP: blood pressure. ABG: Arterial Blood Gases analysis..
As shown in Table 5, patients who achieved the glycemic control target had a lower risk of prolonged hospital stay compared to those who did not achieve it, with a significant association observed in the adjusted analysis (aRR: 0.86; 95% CI: 0.74-0.99). No significant associations were found between glycemic control achievement and other clinical outcomes evaluated. In the sub-analysis of only patients with DM, similar results were found as in the full cohort: glycemic control was associated with a shorter prolonged hospital stay, with a RR of 0.77 (95% CI: 0.67-0.90) and an aRR of 0.79 (95% CI: 0.68-0.93), adjusted for history of neoplasms, lymphocyte count, metabolic state, and hypoxemia.
Table 5. Association between achieving glycemic control and clinical outcomes in patients with hyperglycemia admitted to the emergency department of three Peruvian hospitals.
Outcome, n (%) |
Did not achieve goal (n=400) |
Achieved goal (n=330) |
Total (n=730) |
RR (IC95%) |
p-value |
RRa (IC95%) |
p-value |
Prolonged stay |
242 (60,5) |
171 (51,8) |
413 (56,6) |
0,86 (0,75–0,98) |
0,020 |
0,86 (0,74-0,99) |
0,031 |
Mechanical ventilation |
23 (5,8) |
29 (8,8) |
52 (7,1) |
1,53 (0,90–2,59) |
0,115 |
- |
- |
*AKI with hemodialysis |
18 (4,6) |
13 (3,9) |
31 (4,4) |
0,88 (0,44–1,78) |
0,727 |
- |
- |
Death |
15 (3,8) |
14 (4,2) |
29 (4,0) |
1,13 (0,55–2,31) |
0,735 |
- |
- |
Combined outcome |
51 (12,8) |
45 (13,6) |
96 (13,2) |
1,07 (0,74–1,55) |
0,724 |
- |
- |
RR: Relative Risk. aRR: Adjusted Relative Risk. AKI: Acute Kidney Injury. DM: Diabetes Mellitus.
DISCUSIÓN
Several studies in surgical settings have shown that adequate glycemic control during the perioperative period is associated with better clinical outcomes, including shorter hospital stays. For example, the study by Kurtoglu et al.
12
demonstrated that implementing a glycemic control protocol in patients undergoing major abdominal surgery not only reduced the rate of hyperglycemia but also decreased the time required to reach glycemic values within the target range, which is associated with faster recovery and, presumably, shorter hospitalization. Similarly, in the meta-analysis by Eckert et al.
13
, although no statistical significance was reached in the reduction of ICU stay days, a favorable trend was observed with the use of specialized enteral formulas for glycemic control, suggesting a possible indirect benefit through greater metabolic stability. These findings support those observed in the present study, where early achievement of glycemic control seems to have had a positive impact on reducing prolonged stays. Clinically, these results underline the relevance of establishing early intervention protocols in the emergency department to normalize blood glucose from the moment of admission, aiming to optimize hospital care efficiency and reduce costs associated with unnecessary prolonged hospitalization.
The finding of a reduction in prolonged hospital stays after achieving early glycemic control is also supported by studies conducted in intensive care and other clinical settings. Becker et al.
14
, in a retrospective study in a high-complexity medical ICU, demonstrated that patients with acceptable glycemic control (<180 mg/dL) had a lower probability of experiencing hospital and ICU stays longer than predicted, even after adjusting for severity variables. Similarly, Rady et al.
15
identified that persistently elevated blood glucose was associated with longer mechanical ventilation duration and prolonged stays, particularly in non-diabetic patients, which suggests that sustained metabolic dysfunction may reflect a more intense inflammatory response or greater disease severity. In a different setting, Mozazfia et al.
16
also found that a greater admission glycemic gap (AGG) was related to worse outcomes in neurocritical patients, implying that stress hyperglycemia not corrected early may prolong clinical evolution. The AGG is defined as the difference between the plasma glucose upon admission and the patient's estimated average chronic glucose value, usually calculated from glycated hemoglobin. This indicator helps distinguish acute hyperglycemia from the chronic component and has been proposed as a more precise marker of the metabolic impact of acute stress. Together, these findings suggest that timely glycemic control, beyond its effect on metabolic parameters, could indirectly modulate the progression of acute disease, thereby reducing the duration of hospital stays. Therefore, this study reinforces the need to implement early intervention strategies, even outside critical care units, as the benefits observed in ICUs may be extrapolated, with the advantage of applying them in the earlier stages of care.
Despite the observed benefit in reducing prolonged hospital stays, this study did not find a decrease in mortality, the need for mechanical ventilation, or emergency hemodialysis upon achieving glycemic control in the first 24 hours. This result contrasts with certain studies in surgical settings, where glycemic control has shown associations with a reduction in major complications. For instance, Yang et al.
17
reported that patients with HbA1c ≥6.0% and postoperative blood glucose >200 mg/dL had a fourfold higher risk of postoperative complications following emergency surgery, suggesting that persistent hyperglycemia may be linked to adverse outcomes. However, it is important to note that the strategies and timelines for achieving glycemic control were not detailed in that study, limiting the comparability with the early approach adopted in our work. Additionally, Taylor et al.
18
, through the SUGAR initiative, significantly improved postoperative glycemic control but failed to reduce the incidence of infections or other complications, which coincides with the lack of impact on major clinical events observed in our investigation. Together, these results indicate that glycemic control alone, even if achieved early, may not be sufficient to modify severe clinical outcomes if it is not accompanied by a comprehensive strategy that includes other pathophysiological and contextual determinants. Therefore, although early control is valuable, it should not be overestimated as the sole prognostic measure in acute settings.
In intensive care settings, the relationship between glycemic control and major clinical outcomes has been widely debated, and the findings of the present study — in which early glycemic control did not reduce mortality, mechanical ventilation, or hemodialysis admission — align with multiple pieces of evidence that question the direct clinical benefit of strict glycemic control in critically ill patients. The meta-analysis by Eckert et al.
13
, for example, showed that although specialized formulas for glycemic control in critically ill patients reduced blood glucose levels and insulin requirements, there was no significant impact on mortality, mechanical ventilation duration, or ICU stay days. Similarly, the multicenter trial conducted by Agus et al.
19
in critically ill children found no differences in mortality, ventilation, or ICU-free days between strict glycemic control (80–110 mg/dL) and moderate control (150–180 mg/dL), but there was an increase in severe hypoglycemia in the intensive intervention group, questioning the risk-benefit profile of aggressive intervention. Likewise, studies by Rady et al. and Becker et al.
14
,
15
suggest that while there is an association between hyperglycemia and mortality, this relationship is modulated by multiple individual factors such as the underlying diagnosis, inflammatory response, use of steroids or catecholamines, and the presence or absence of pre-existing diabetes. These findings reinforce the idea that early glycemic control in itself is not a sufficient isolated tool to modify life prognosis or avoid advanced life support, and its impact may be dependent on the patient's risk profile, the etiology of the acute illness, and the pathophysiological moment at which intervention occurs. Therefore, in the emergency setting, where clinical heterogeneity is considerable, interventions must be personalized.
The results of the present study also echo investigations conducted in other clinical settings, where the impact of glycemic control on mortality and other adverse outcomes has been variable and, in many cases, limited. For example, in patients with acute cerebrovascular disease, both the meta-analysis by Wu et al.
20
, and the SHINE study analysis
21
found no significant benefits from intensive glycemic control on mortality, functionality at 90 days, or event recurrence, although there was an increased risk of severe hypoglycemia. Similarly, in hospitalized COVID-19 patients, Klonoff et al.
22
reported that sustained hyperglycemia on days 2 or 3 was associated with higher mortality only in non-critical patients, while in the ICU, this relationship was not significant after the second day, suggesting a narrow time window for glycemic control to impact prognosis. In neurocritical patients, Mozazfia et al.
16
found that a higher AGG was associated with higher mortality, indicating that beyond the absolute value of glucose, the magnitude of the acute glycemic imbalance relative to prior chronic control may be more relevant. Additionally, studies such as those by Li and Yuan
23
, in severe coronary disease, and the meta-analysis by Crabtree et al.
24
in older and frail adults, highlight that the intensity of glycemic control must be carefully individualized, as excessively strict control may be harmful in certain groups. In light of this evidence, the results of the present study reinforce the notion that early glycemic control is an intervention with relevant logistical and metabolic potential — such as in reducing hospital stay — but it should not be assumed as a universally effective strategy for preventing major events in all clinical contexts. Therefore, it is recommended to direct emergency glycemic management toward early but safe control, avoiding extremes, and considering other pathophysiological and prognostic variables that may modulate the expected clinical benefit in each patient.
This study has limitations inherent to its retrospective design, such as dependence on incomplete or heterogeneous clinical records between hospitals, which could have introduced information bias. The use of non-probabilistic convenience sampling limits the generalizability of the findings to other populations. Furthermore, some potentially influential clinical variables, such as baseline functional status or degree of dehydration, were not considered. Finally, although multivariate models were applied, residual confounding factors that were not controlled for cannot be ruled out..
CONCLUSIÓN
The findings of this study suggest that achieving glycemic control within the first 24 hours of admission to the emergency department is associated with a reduction in prolonged hospital stays, representing a relevant clinical and logistical benefit in the care of patients with acute hyperglycemia. However, this early control did not demonstrate a significant impact on mortality or the occurrence of critical outcomes such as the need for mechanical ventilation or emergency hemodialysis. These results indicate that while timely glycemic control may contribute to a more favorable and efficient progression, its effect on major clinical outcomes appears to be conditioned by multiple factors, so it is recommended to implement it as part of a comprehensive and personalized approach to the management of acute patients.