Best evidence topic report: Does Baseline Serum Lactate Predict Outcomes in Diabetic Ketoacidosis Patients?

Report By: Mostafa Alavi-Moghaddam - Full Professor of Medical School Department of Emergency Medicine
Search checked by Ramin Zeinodini - Research Associate
Institution: Emergency Medicine Department, Imam Hossein Hospital, Shahid Beheshti University of Medical Sciences, Tehran, Iran
Date Submitted: 27th May 2025
Last Modified: 2nd June 2025
Status: Blue (submitted but not checked)

Three Part Question

In [patients with diabetic ketoacidosis (DKA)], does [a higher baseline serum lactate level compared to a lower level] predict [worse outcomes, such as increased hospital or intensive care unit length of stay, morbidity, and mortality]?

Clinical Scenario

A young man with type 1 diabetes presented to the emergency department (ED) with disorientation and was diagnosed with DKA. Despite his elevated baseline serum lactate, his condition improved relatively quickly. This makes you wonder whether high baseline serum lactate predicts outcomes in DKA patients.

Search Strategy

PubMed up to 4/19/2025: (((("Diabetic Ketoacidosis"[Mesh]) OR (Diabetic Ketoacidos*[Title/Abstract] OR Diabetic Ketos*[Title/Abstract] OR Diabetic Acidos*[Title/Abstract])) AND (("Lactates"[Mesh]) OR (Lactate*[Title/Abstract] OR Lactic Acid*[Title/Abstract] OR "2-Hydroxypropanoic Acid*"[Title/Abstract] OR "2 Hydroxypropionic Acid*"[Title/Abstract] OR "2 Hydroxypropanoic Acid*"[Title/Abstract] OR "2-Hydroxypropionic Acid*"[Title/Abstract] OR "L Lactic Acid*"[Title/Abstract] OR "L-Lactic Acid*"[Title/Abstract] OR "D Lactic Acid*"[Title/Abstract] OR "D-Lactic Acid*"[Title/Abstract]))) AND (outcome[Title/Abstract] OR prognos*[Title/Abstract] OR mortality[Title/Abstract] OR morbidity[Title/Abstract] OR death[Title/Abstract] OR complication[Title/Abstract] OR surviv*[Title/Abstract] OR "Intensive care unit"[Title/Abstract] OR critical[Title/Abstract])) AND (predict*[Title/Abstract] OR associat*[Title/Abstract] OR correlat*[Title/Abstract] OR 'risk factor'[Title/Abstract] OR relat*[Title/Abstract]) AND ((humans[Filter]) AND (english[Filter]))
Embase up to 4/19/2025: ('lactic acid'/exp OR '2 hydroxypropanoic acid':ab,ti OR '2 hydroxypropionic acid':ab,ti OR 'l lactate':ab,ti OR 'l lactic acid':ab,ti OR 'l-lactate':ab,ti OR 'l-lactic acid':ab,ti OR 'd lactate':ab,ti OR 'd lactic acid':ab,ti OR 'd-lactate':ab,ti OR 'd-lactic acid':ab,ti OR 'lactate':ab,ti OR 'lactic acid':ab,ti) AND ('diabetic ketoacidosis'/exp OR 'diabetic ketoacidosis' OR 'diabetic acidosis':ab,ti OR 'diabetic keto-acidosis':ab,ti OR 'diabetic ketoacidoses':ab,ti OR 'diabetic ketoacidosis':ab,ti OR 'diabetic ketosis':ab,ti OR 'diabetic ketoses':ab,ti OR 'diabetic acidoses':ab,ti) AND ('outcome':ab,ti OR 'intensive care unit':ab,ti OR 'prognos*':ab,ti OR 'surviv*':ab,ti OR 'mortality':ab,ti OR 'morbidity':ab,ti OR 'death':ab,ti OR 'complication':ab,ti OR 'critical':ab,ti) AND ('predict*':ab,ti OR 'associat*':ab,ti OR 'correlat*':ab,ti OR 'risk factor':ab,ti OR 'relat*':ab,ti) AND [humans]/lim AND [english]/lim AND [embase]/lim
Cochrane up to 4/19/2025:
#1 MeSH descriptor: [Lactic Acid] explode all trees
#2 MeSH descriptor: [Diabetic Ketoacidosis] explode all trees
#3 ("Lactate" OR "Lactic Acid" OR "2-Hydroxypropanoic Acid" OR "2 Hydroxypropionic Acid" OR "2 Hydroxypropanoic Acid" OR "2-Hydroxypropionic Acid" OR "L Lactic Acid" OR "L-Lactic Acid" OR "D Lactic Acid" OR "D-Lactic Acid"):ti,ab,kw
#4 (diabet*):ti,ab,kw AND (Ketoacidos* OR Ketos* OR Acidos*):ti,ab,kw
#5 ("outcome" OR prognos* OR "mortality" OR "morbidity" OR "death" OR "complication" OR surviv* OR "Intensive care unit" OR "critical"):ti,ab,kw
#6 (predict* OR associat* OR correlat* OR "risk factor" OR relat*):ti,ab,kw
#7 #1 OR #3
#8 #2 OR #4
#9 #7 AND #8 AND #5 AND #6

Search Outcome

A total of 377 papers were found, of which 327 were identified after removing duplicates. Of these, 303 papers were excluded based on title and abstract screening. The remaining 24 full-text articles were screened and assessed for eligibility. Finally, 15 studies were enrolled and provided the best evidence to answer the three-part question (Figure 1).

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Cox 2012 USA	68 ED adult patients with DKA, categorized by serum lactate (mmol/L): <4 (n=41) ≥4 (n=27)	Retrospective cohort study	ICU LOS, Mortality	• ICU LOS: p=0.18 (ns diff.) • Mortality: p=1.0 (ns diff.)	Measured serum lactate only in 52% of DKA cases, low mortality, potential missed interactions, retrospective design, small sample size, single-center study
Suwarto 2014 Indonesia	60 ED adult patients with DKA, categorized by serum lactate (mmol/L): <4 (n=43) ≥4 (n=17)	Prospective cohort study	Five-day mortality	Five-day mortality: • For serum lactate ≥4 mmol/L (vs <4 mmol/L): HR=3.09 (95% CI: 1.36–7.05), p-value not reported	Unexamined comorbidities, unclear causes of altered mental status, small sample size, single-center study
Siregar 2018 Indonesia	260 adult DKA patients with serum lactate data, categorized by serum lactate (mmol/L): <2 (n=138) 2-4 (n=69) >4 (n=53)	Retrospective cohort study	Seventy-two-hour mortality	Seventy-two-hour mortality: • For serum lactate 2–4 mmol/L (vs <2 mmol/L): HR=3.117 (95% CI: 1.609–6.037), p=0.001 • For serum lactate >4 mmol/L (vs <2 mmol/L): HR=5.585 (95% CI: 2.966–10.519), p<0.001	Predominantly type 2 diabetes (90%), retrospective design, single-center study
Lee 2019 Republic of Korea	274 DKA patients with serum lactate data, categorized by serum lactate (mmol/L): ≤5 (n=246) >5 (n=28)	Retrospective cohort study	In-hospital mortality	In-hospital mortality: For serum lactate >5 mmol/L (vs ≤5 mmol/L): • Adjusted OR=1.275 (95% CI: 1.132–1.435), p<0.01 (dir. corr.)	Excluded euglycemic DKA cases, retrospective design, single-population
Cully 2020	92 ED paediatric patients (113 encounters) with severe DKA, categorized by serum lactate (mmol/L): <2.5 (n=41) ≥2.5 (n=72)	Retrospective cohort study	Hospital LOS, Paediatric ICU LOS, Morbidity, Mortality	• Hospital LOS: p=0.426 (ns diff.) • Paediatric ICU LOS: p=0.321 (ns diff.) • Morbidity: p=0.552 (ns diff.) • Mortality: p=1.000 (ns diff.)	Low morbidity and mortality, retrospective design, small sample size, single-center study
Unal 2020 Turkey	56 paediatric ICU patients with DKA (aged 2–18 years), categorized by serum lactate (mmol/L): <2 (n=18) >2 (n=38)	Retrospective cohort study	DKA resolution time, Mortality	• DKA resolution time: p=0.26 (ns diff.) • Lactate levels: Negative correlation with DKA resolution time (r=-0.33, p=0.01) • No mortality observed.	No mortality data, excluded confounding conditions, retrospective design, small sample size, single-center study
Morgan 2021 Australia	25 ICU patients with DKA, categorized by serum lactate (mmol/L): ≤2 (n=3) >2 (n=22)	Retrospective cohort study	Hospital LOS, ICU LOS	• Hospital LOS: p=0.579 (ns diff.) • ICU LOS: p=0.405 (ns diff.)	Possible missed cases due to exclusion criteria, retrospective design, small sample size, single-center study
Bhat 2021 India	62 tertiary care hospital patients with DKA, categorized by serum lactate (mmol/L). <2.5 (n=27) ≥2.5 (n=35)	Prospective cohort study	Hospital LOS, Morbidity, Mortality	• Hospital LOS: p=0.579 (ns diff.) • No mortality or significant morbidity observed.	No mortality data, unclear metformin association, small sample size, single-center study
Liu 2021 China	107 paediatric ICU patients with DKA, categorized by serum (mmol/L): ≤2 (n=46) >2 (n=61)	Retrospective cohort study	DKA resolution time, Hospital LOS	• DKA resolution time: r=0.143, p=0.143 (ns corr.) • Hospital LOS: r=0.154, p=0.115 (ns corr.)	Lack of mortality data, retrospective design, small sample size, single-center study
Taşkın 2021 Turkey	40 adult ICU patients with DKA, categorized by serum lactate (mmol/L): <4 (n=24) ≥4 (n=16)	Retrospective cohort study	Serum lactate, Hospital LOS, ICU LOS, Mortality	• Serum lactate: Significantly higher in non-survivors at both hospital (p=0.017) and ICU admission (p=0.010) compared to survivors. • Hospital LOS: p=0.274 (ns diff.) • ICU LOS: p=0.966 (ns diff.) • Mortality: p=0.195 (ns diff.)	Unclear diabetes type differentiation, retrospective design, small sample size, single-center study
Kart 2023 Turkey	102 ED patients with DKA, categorized by mortality status.	Retrospective cohort study	Serum lactate (hour 1)	• Serum lactate (hour 1): Significantly higher in non-survivors compared to survivors (p=0.024).	Retrospective design, small sample size, single-center study
Kumar 2023 New Zealand	9061 adult ICU patients with DKA, categorized by serum lactate (mmol/L): <2 (n=5674) ≥2 (n=3405)	Retrospective, multicenter cohort study	Hospital LOS, ICU LOS, In-hospital mortality	For serum lactate ≥2 mmol/L (Ref: <2 mmol/L, persisted only in T2D, not T1D, after separation): • Hospital LOS: GMR=1.063 (95% CI: 1.025–1.103), p=0.001 (dir. corr.) • ICU LOS: GMR=1.057 (95% CI: 1.026–1.090), p<0.001 (dir. corr.) • In-hospital mortality: OR=1.785 (95% CI: 1.122–2.841), p=0.014 (dir. corr.)	Selective ICU data, missing serum lactate data, retrospective design
Ozel 2024 Turkey	112 paediatric T1D patients with DKA (aged 7 months to 17 years), categorized by serum lactate (mmol/L): <2.5 (n=27) ≥2.5 (n=35)	Retrospective cohort study	DKA resolution time, Paediatric ICU LOS	• DKA resolution time: p=0.746, r=0.031 (ns corr.) • PICU LOS: p=0.109, r=0.152 (ns corr.)	Lack of mortality data, excluded patients with confounding conditions, retrospective design, small sample size, single-center study
Ruttinger 2025 Austria	523 patients with DKA, categorized by 7-day mortality status.	Retrospective cohort study	7-day mortality	• Serum lactate: Significantly higher (p<0.001) in non-survivors (6.3 mmol/l) compared to survivors (3.4 mmol/l)	Not distinguished diabetes types (T1D vs. T2D), missing lactate data for some patients, retrospective design
Söbü 2025 Turkey	56 pediatric ED patients with DKA (aged 35-210 months), categorized by serum lactate (mmol/L): <2 (n=29) ≥2 (n=27)	Retrospective cohort study	Serum lactate, Hospital LOS, ICU LOS	• Hospital LOS: p=0.325 (ns diff.) • ICU LOS: p=0.616 (ns diff.)	No morbidity or mortality data, retrospective design, small sample size

Comment(s)

The emergency physicians who manage DKA in the ED must decide whether the patient requires advanced monitoring or intensive care. The question is: Does baseline serum lactate predict worse outcomes? Out of 15 related studies (1-15), nine studies (1, 3, 6-11, 15) with a total of 714 participants, including all paediatric studies (1, 3, 7, 10, 11), found no significant association between higher baseline serum lactate levels and worse outcomes. On the other hand, among the remaining six studies (2, 4, 5, 12-14) supporting this association, a large population study (4) with 9,061 participants revealed that this association was only valid for type 2 diabetic patients. Taken together, these results suggest that baseline serum lactate levels may predict worse outcomes in DKA, specifically in type 2 diabetic adults. However, further studies across diverse populations, considering the type of diabetes, are required to confirm these findings.

Clinical Bottom Line

Baseline serum lactate levels may predict outcomes in type 2 diabetic adults with DKA, but further studies across diverse populations, considering the type of diabetes, are required.