Acid–Base Phenotypes in Sedative Toxidrome: Characterisation, Agent‐Specific Patterns and Clinical Associations

ABSTRACT

Objective

Acid–base patterns and their association with clinical outcomes have not been characterised using venous‐specific reference ranges in a sedative‐specific cohort. We aimed to characterise acid–base phenotypes in sedative toxidrome using venous‐specific thresholds, examine agent‐specific associations and evaluate relationships to adverse outcomes.

Methods

Retrospective cohort study of 374 consecutive patients referred to a toxicology service with sedative poisoning between November 2022 and November 2023. VBG data (pH, pCO₂, HCO₃ and lactate) were classified into acid–base phenotypes using venous‐specific reference ranges (venous respiratory acidosis: pH < 7.31, pCO₂ > 51 mmHg; hypercapnia with preserved pH: pH 7.31–7.41, pCO₂ > 51 mmHg). Outcomes of interest were endotracheal intubation, ICU admission and in‐hospital complications. Associations were assessed using chi‐squared tests, Mann–Whitney U tests and ROC curve analysis.

Results

VBG was obtained in 330 of 374 patients (88.2%). Venous respiratory acidosis occurred in 54 patients (16.4%), hypercapnia with preserved pH in 81 (24.5%) and normal venous pH in 153 (46.4%). Opioid exposures had the highest hypercapnic burden (59%) compared with anti‐psychotic exposures (20%; p < 0.001). Venous respiratory acidosis was associated with ICU admission (40.7% vs. 17.0%, p = 0.001) and complications (64.8% vs. 32.7%, p < 0.001). Among hypercapnic patients, pCO₂ ≥ 65 mmHg was associated with ICU admission in 50.0%. Venous pH had the strongest discriminative performance across outcomes (AUC 0.631–0.683).

Conclusions

Venous respiratory acidosis identifies a high‐risk subgroup. Hypercapnic burden varies substantially by agent class. These findings support routine early VBG measurement in sedative overdose and highlight the importance of venous‐specific thresholds when interpreting point‐of‐care acid–base results.

1. Introduction

Sedative drug overdose is an increasing contributor to emergency department (ED) presentations and intensive care unit (ICU) admissions, accounting for a significant proportion of medical indications for intubation in Australia and New Zealand [1, 2]. While the clinical and physiological risk factors associated with deterioration in sedative overdose have been investigated, there is limited data on acid–base disturbance patterns and their agent‐specific characteristics. Abnormal acid–base findings are often used in the ED to facilitate early decision‐making regarding treatment, including the need for intubation.

Venous blood gas (VBG) analysis provides rapid, point‐of‐care acid–base assessment and is routinely performed in ED patients to assist with diagnosis, risk stratification and clinical progression. Previous work from our group demonstrated that venous pH is a strong independent predictor of intubation, ICU admission and complications in this population, with each 0.1‐unit decrease in pH associated with a 3%–4% increase in the odds of adverse outcomes [3]. However, that analysis treated pH as a continuous variable and did not characterise the underlying patterns of acid–base disturbance nor their relationship to specific agent classes.

Different sedative agents produce acid–base derangement through distinct mechanisms. Opioids and gamma‐hydroxybutyrate (GHB) primarily cause respiratory centre depression with consequent hypoventilation and respiratory acidosis [4, 5]. In contrast, anti‐psychotics and anti‐depressants may produce less severe respiratory depression, and ethanol causes a characteristic Type B lactic acidosis through inhibition of gluconeogenesis that may be independent of ventilatory compromise [6]. Understanding these agent‐specific patterns could inform tailored monitoring strategies and early escalation thresholds as well as assisting with overall risk assessment.

Prior studies in poisoned patients have largely examined lactate as a mortality predictor or compared individual VBG parameters without characterising discrete patterns of acid–base disturbance (acid–base phenotypes) [7, 8]. The relationship between acid–base disturbance patterns, specific sedative agent classes and clinical outcomes has not been systematically examined in a sedative‐specific toxidrome cohort. This study addresses this gap using a prospectively maintained, consecutive toxicology cohort.

2. Methods

2.1. Study Design and Cohort

This was a pre‐specified secondary analysis of a retrospective cohort study of patients referred to the Western Sydney Toxicology Service between November 2022 and November 2023. This service covers a large local government area of western Sydney with a population of 1.2 million and four EDs and two ICUs. The toxicology service provides 24/7 specialist cover and maintains a prospectively collected database of all admissions and consultations since 2010. Patient information was extracted from the electronic medical record (Oracle Health Millennium, Kansas City, MO). Patients were excluded if altered consciousness was attributed to non‐toxicological causes, if the exposure did not involve a sedative agent, or if clinical data were insufficient for outcome assessment. Only the index presentation was included for each patient. The study was approved by the Western Sydney Local Health District Human Research Ethics and Governance Office as a secondary analysis of previously approved data.

2.2. Acid–Base Classification

Patients with complete VBG data (pH, pCO₂ and HCO₃) were available from the initial ED presentation pre‐intubation or airway intervention. These were classified using venous‐specific reference ranges, as VBG values differ systematically from arterial values (venous pH approximately 0.03–0.05 units lower; venous pCO₂ approximately 4–6 mmHg higher) [9]. Venous respiratory acidosis was defined as pH < 7.31 with pCO₂ > 51 mmHg and HCO₃ ≥ 22 mmol/L. A separate category of hypercapnia with preserved pH was defined as pH 7.31–7.41 with pCO₂ > 51 mmHg, reflecting elevated pCO₂ without frank venous acidaemia. Metabolic acidosis was defined as pH < 7.31 with HCO₃ < 22 mmol/L regardless of pCO₂. Mixed acidosis required pH < 7.31, pCO₂ > 51 mmHg and HCO₃ < 22 mmol/L simultaneously. Normal venous pH was defined as pH 7.31–7.41 with pCO₂ ≤ 51 mmHg and HCO₃ ≥ 22 mmol/L. Alkalosis was defined as pH > 7.41. Elevated lactate was defined as > 2.0 mmol/L and was analysed as a concurrent finding rather than a primary classification criterion. Patients without complete VBG data were classified as no VBG obtained and were included in outcome analyses.

2.3. Outcomes

The outcomes of interest were endotracheal intubation, ICU admission and any in‐hospital complication, as defined in the primary publication [3]. Hospital admission alone was not considered clinical deterioration. Complication was defined as the occurrence of any adverse clinical event during the index admission, consistent with the definition used in the primary publication and including aspiration pneumonia, seizure, cardiac arrest, shock, acute kidney injury, rhabdomyolysis, sepsis and other toxidrome‐specific complications. A complete list of complications is provided in Table S1.

2.4. Statistical Analysis

Continuous variables were summarised as medians with interquartile ranges (IQR) and compared using the Mann–Whitney U test. Categorical variables were compared using chi‐squared or Fisher's exact tests as appropriate. Where appropriate, absolute risk differences with 95% confidence intervals were calculated to supplement p‐values and aid clinical interpretation. Outcome rates were compared across acid–base categories using chi‐squared with Bonferroni correction for multiple comparisons where appropriate. The discriminative performance of individual VBG parameters (pH, pCO₂, HCO₃ and lactate) for each outcome was assessed using the area under the receiver operating characteristic curve (AUC), with 95% confidence intervals derived by bootstrap resampling (1000 iterations). Statistical analyses were performed using Python 3 (Python Software Foundation, Wilmington, DE, the USA). A two‐sided p‐value < 0.05 was considered statistically significant.

3. Results

3.1. VBG Data Availability and Cohort Characteristics

During the study period, 374 consecutive patients with suspected or confirmed sedative poisoning were included, of whom complete VBG data were available for 330 (88.2%). The 44 patients (11.8%) in whom VBG was not obtained had 0% intubation and 0% ICU admission. All VBGs were obtained prior to any airway intervention. Among those with a VBG taken, the median pH was 7.350 (IQR 7.310–7.390), median pCO₂ 49.0 mmHg (IQR 43.0–56.0), median HCO₃ 27.0 mmol/L (IQR 25.0–29.0), and median lactate 1.30 mmol/L (IQR 0.90–2.00). Elevated lactate (> 2.0 mmol/L) was present in 74 patients (22.4% of those with lactate data).

3.2. Acid–Base Phenotype Distribution

The distribution of acid–base phenotypes and associated clinical outcomes are presented in Table 1. Among the 330 patients with VBG data, normal pH was the most frequent finding (153, 46.4%), followed by hypercapnia with preserved pH (81, 24.5%) and venous respiratory acidosis (54, 16.4%). Metabolic acidosis was identified in four patients (1.2%) and mixed acidosis in three patients (0.9%). Thirty‐five patients (10.6%) had pH > 7.41, meeting the venous alkalosis threshold; the majority of these patients (median pH 7.430, IQR 7.420–7.450) would be classified as normal by conventional arterial criteria and their outcomes were similar to the normal venous pH group (intubation 11.4%, ICU 11.4%).

TABLE 1.

Acid–base phenotype distribution and clinical outcomes.

Acid–base pattern	N (%)	Intubation n (%)	ICU admission n (%)	Any complication n (%)	Median pH (IQR)
Normal venous pH	153 (46.4%)	22 (14.4%)	26 (17%)	50 (32.7%)	7.370 (7.340–7.390)
Hypercapnia, preserved pH	81 (24.5%)	6 (7.4%)	10 (12.3%)	22 (27.2%)	7.330 (7.320–7.360)
Venous respiratory acidosis	54 (16.4%)	15 (27.8%)	22 (40.7%)	35 (64.8%)	7.280 (7.240–7.300)
Metabolic acidosis	4 (1.2%)	1 (25.0%)	2 (50.0%)	2 (50.0%)	7.195 (7.162–7.230)
Mixed acidosis	3 (0.9%)	3 (100%)	3 (100%)	3 (100%)	7.140 (7.125–7.175)
Alkalosis a	35 (10.6%)	4 (11.4%)	4 (11.4%)	10 (28.6%)	7.430 (7.420–7.450)
No VBG obtained	44 (—)	0 (0%)	0 (0%)	7 (15.9%)	—

Note: Outcomes presented as n (%). Venous respiratory acidosis defined as pH < 7.31 with pCO₂ > 51 mmHg and HCO₃ ≥ 22 mmol/L. Hypercapnia with preserved pH defined as pH 7.31–7.41 with pCO₂ > 51 mmHg. Metabolic and mixed acidosis groups are small (n = four and n = three, respectively) and findings are descriptive only. Reference category for comparisons is normal venous pH.

Abbreviation: VBG = venous blood gas.

^a Patients with pH 7.41–7.45 meet the venous alkalosis threshold but would be classified as normal by conventional arterial criteria; their clinical outcomes were similar to the normal venous pH group.

3.3. Outcomes by Acid–Base Phenotype

Venous respiratory acidosis was independently associated with higher rates of ICU admission (40.7% vs. 17.0% for normal venous pH, absolute difference 23.7%, 95% CI 9.4%–38.1%, p = 0.001) any complication (64.8% vs. 32.7%, absolute difference 32.1%, 95% CI 17.4%–46.9%, p < 0.001) and intubation (27.8% vs. 14.4%, absolute difference 13.4%, 95% CI 0.2%–26.6%, p = 0.045). The hypercapnia with preserved pH group had outcome rates similar to or lower than the normal venous pH group (intubation 7.4%, ICU 12.3%, complication 27.2%). Outcome rates for metabolic and mixed acidosis groups are presented in Table 1; these groups comprised three and four patients respectively and findings are descriptive only. Patients with normal venous pH had the lowest outcome rates across all three endpoints.

3.4. Venous Blood Gas Parameters by Outcome Group

Across all three outcomes, pH was the VBG parameter most consistently associated with clinical deterioration. Patients who were intubated had a significantly lower median pH (7.330, IQR 7.275–7.370) compared with those who were not (7.360, IQR 7.320–7.390; difference −0.030, p = 0.003). Patients admitted to ICU had lower median pH (7.320, IQR 7.255–7.365 vs. 7.360, IQR 7.320–7.390; difference −0.040, p < 0.001) and higher median pCO₂ (51.0 vs. 48.0 mmHg; difference +3.0 mmHg, p = 0.018). HCO₃ did not differ significantly between outcome groups for any endpoint.

3.5. Performance of VBG Parameters

AUC values for individual VBG parameters across the three outcome measures are presented in Table 2. Venous pH demonstrated the highest discriminative performance across all outcomes (AUC 0.631–0.683). pCO₂, HCO₃ and lactate each showed weaker discrimination when considered in isolation (AUC 0.546–0.600).

TABLE 2.

Discriminative performance of individual VBG parameters (AUC with 95% CI).

Outcome	pH AUC	95% CI	pCO2 AUC	95% CI	Lactate AUC
Intubation	0.631	0.542–0.720	0.556	0.459–0.645	0.563 (0.472–0.652)
ICU admission	0.683	0.599–0.769	0.593	0.507–0.673	0.600 (0.521–0.682)
Any complication	0.622	0.554–0.690	0.560	0.491–0.623	0.574 (0.514–0.638)

Note: AUCs calculated using the inverse of pH and HCO₃ (lower = worse) and direct values of pCO₂ and lactate (higher = worse). 95% CIs derived by bootstrap resampling (n = 1000). Lactate AUC presented with CI in the final column. AUCs calculated using continuous VBG values independently of categorical acid–base classification, using the inverse of pH and HCO₃ (lower = worse) and direct values of pCO₂ and lactate (higher = worse). 95% CIs derived by bootstrap resampling (n = 1000). Lactate AUC presented with CI in the final column.

3.6. Agent‐Specific Acid–Base Phenotypes

Venous respiratory acidosis and hypercapnia with preserved pH prevalence differed across agent classes (Table 3, Figure 1). Opioid exposures had the highest rates of venous respiratory acidosis (32%), followed by non‐benzodiazepine GABA‐ergic agents (22%) and ethanol (22%), compared with anti‐psychotic exposures (six patients, 6%). Benzodiazepine and GHB exposures demonstrated intermediate venous respiratory acidosis rates (17% for both). When hypercapnia with preserved pH was included alongside venous respiratory acidosis, opioid exposures had the highest combined hypercapnic burden (58%), compared with anti‐psychotics (20%).

TABLE 3.

Venous blood gas parameters and acid–base phenotypes by agent class.

Agent class	N	Median pH (IQR)	Median pCO2 (IQR) mmHg	Venous respiratory acidosis n (%)	Hypercapnia, preserved pH n (%)	Any hypercapnia	Elevated lactate n (%)
Overall	330	7.350 (7.310–7.390)	49.0 (43.0–56.0)	54 (16%)	81 (25%)	135 (41%)	74 (23%)
Opioid	99	7.330 (7.295–7.375)	54.0 (47.0–62.0)	32 (32%)	26 (26%)	58 (59%)	25 (25%)
GHB	72	7.340 (7.300–7.380)	50.0 (44.0–57.0)	12 (17%)	20 (28%)	32 (44%)	11 (15%)
Benzodiazepine	108	7.350 (7.310–7.390)	51.0 (44.0–57.0)	18 (17%)	36 (33%)	54 (50%)	20 (19%)
Non‐benzo GABA‐ergic	23	7.350 (7.320–7.385)	52.0 (44.0–57.0)	5 (22%)	7 (30%)	12 (52%)	5 (22%)
Anti‐depressant	42	7.365 (7.330–7.395)	46.5 (42.5–52.5)	3 (7%)	9 (21%)	15 (36%)	14 (33%)
Anti‐psychotic	64	7.380 (7.350–7.400)	44.5 (40.5–49.0)	4 (6%)	9 (14%)	13 (20%)	13 (20%)
Ethanol	32	7.360 (7.330–7.390)	47.0 (43.0–52.0)	7 (22%)	5 (16%)	12 (38%)	15 (47%)

Note: Venous respiratory acidosis (pH < 7.31, pCO₂ > 51 mmHg, HCO₃ ≥ 22 mmol/L). Hypercapnia with preserved pH defined as pH 7.31–7.41 with pCO₂ > 51 mmHg. Non‐benzo GABA‐ergic agents include pregabalin, gabapentin and Z‐drugs. Elevated lactate defined as > 2.0 mmol/L. Patients may be represented in more than one agent class row due to polypharmacy. Bolded to highlight the key result.

Abbreviation: GHB = γ‐hydroxybutyrate.

FIGURE 1.

Acid–base phenotype distribution, clinical outcomes and agent‐specific hypercapnia profiles in sedative overdose. (A) Distribution of acid–base phenotypes among 330 patients with complete VBG data, classified using venous‐specific reference ranges. Patients with pH 7.41–7.45 meet the venous alkalosis threshold but would be classified as normal by conventional arterial criteria. (B) Clinical outcome rates (intubation, ICU admission, any complication) by acid–base phenotype. Bracket indicates p = 0.045 for intubation comparison between venous respiratory acidosis and normal venous pH groups. (C) Prevalence of any hypercapnia (venous respiratory acidosis plus hypercapnia with preserved pH, stacked) by agent class, ordered by total hypercapnic burden. Dark red segments represent venous respiratory acidosis (pH < 7.31, pCO₂ > 51 mmHg); light blue segments represent hypercapnia with preserved pH (pH 7.31–7.41, pCO₂ > 51 mmHg). Dashed vertical line indicates overall cohort hypercapnia prevalence (41%). GHB = γ‐hydroxybutyrate; non‐benzo GABA = non‐benzodiazepine GABA‐ergic agents (pregabalin, gabapentin and Z‐drugs); VRA = venous respiratory acidosis.

Lactate profiles also differed by agent class (Figure 2). Prior work demonstrated that benzodiazepine co‐ingestion was inversely associated with hyperlactataemia in a mixed overdose cohort [7]. Our data confirm this finding in a sedative‐specific cohort. Our study further demonstrates that GHB also produces low lactate levels (median 1.20 mmol/L, 15% elevated), consistent with its GABAergic mechanism and absence of sympathomimetic or metabolic toxicity. In contrast, ethanol produced elevated lactate in nearly half of presentations (47%, median 1.95 mmol/L), significantly higher than benzodiazepines (median 1.20 mmol/L; p = 0.002), reflecting its well‐described inhibition of hepatic gluconeogenesis and contribution to Type B lactic acidosis [6]. Anti‐depressants were also associated with elevated lactate (33%, median 1.55 mmol/L; p = 0.028 vs. benzodiazepines), potentially reflecting drug‐induced mitochondrial dysfunction or seizure‐related lactate production. These findings suggest that lactate interpretation should be contextualised by the likely causative agent class.

FIGURE 2.

Acid–base severity and lactate profiles. (A) Clinical outcome rates by pCO₂ severity stratum among all hypercapnic patients (venous respiratory acidosis or hypercapnia with preserved pH combined, n = 135). (B) Lactate distribution by agent class. Red dashed line indicates the elevated lactate threshold (2.0 mmol/L). Asterisks indicate statistically significant difference versus benzodiazepines (*p < 0.05, **p < 0.01).

3.7. pCO₂ Severity Gradient in Respiratory Acidosis

Among patients with any degree of hypercapnia (venous respiratory acidosis or hypercapnia with preserved pH combined, n = 135), pCO₂ severity was associated with a graded increase in adverse outcome rates. Intubation rates were 8.9%, 14.7% and 31.8% for pCO₂ strata of 45–54, 55–64 and ≥ 65 mmHg respectively. ICU admission rates followed the same gradient (17.8%, 19.1% and 50.0%).

3.8. Concurrent Respiratory Acidosis and Elevated Lactate

The combination of venous respiratory acidosis and elevated lactate (pH < 7.31, pCO₂ > 51 mmHg and lactate > 2.0 mmol/L) was present in 15 patients (4.0% of the total cohort). ICU admission occurred in 53.3% of this group compared with 36.8% of those with venous respiratory acidosis and normal lactate, though the small group size limits interpretation of this comparison.

3.9. Polypharmacy and Acid–Base Phenotypes

The majority of patients presented with single‐agent sedative exposures (226, 60.4%), with two‐agent co‐ingestions in 95 patients (25.4%) and polypharmacy, defined as ingestion of more than two agents, in 53 (14.2%). The number of agents ingested was not associated with acid–base phenotype—venous respiratory acidosis rates were similar in patients with and without polypharmacy (20.0% vs. 15.7%, p = 0.595), and the number of agents showed no significant correlation with venous pH (Spearman's r = 0.026, p = 0.638), pCO₂ (r = −0.003, p = 0.962) or HCO₃ (r = 0.030, p = 0.589). Within the venous respiratory acidosis subgroup (n = 54), 10 patients (18.5%) had polypharmacy; outcome rates in this subgroup are presented descriptively in Table 1 given the small numbers.

4. Discussion

This study systematically characterises acid–base phenotypes in a consecutive sedative toxidrome cohort and examines their relationships with agent class and clinical outcomes. We note three key findings. First, venous respiratory acidosis—defined using venous‐specific reference ranges—occurred in 16.4% of patients with VBG data and was associated with substantially higher rates of ICU admission and complications than either normal venous pH or hypercapnia with preserved pH. Second, the prevalence of venous respiratory acidosis and hypercapnia varied by agent class, being the most common in opioid exposures and least common in anti‐psychotic exposures. Third, among all hypercapnic patients, a pCO₂ ≥ 65 mmHg identified a subgroup with ICU admission rates approaching 50%.

The distribution of acid–base disturbance patterns across agent classes aligns with the fundamental pharmacology of sedative toxidrome. Opioids suppress ventilatory drive through activation of μ‐opioid receptors in the brainstem, while GHB causes profound GABA‐mediated CNS depression with associated hypoventilation [4, 5]. The finding that opioids had the highest combined hypercapnic burden (venous respiratory acidosis or hypercapnia with preserved pH, 58%) compared with anti‐psychotic exposures (20%) is consistent with these mechanistic differences. Anti‐psychotics produce sedation through dopamine and histamine receptor antagonism but central respiratory control is preserved to a greater extent at clinically encountered doses [10]. These agent‐specific patterns suggest that patients presenting with opioid toxicity may be at higher physiological risk of respiratory compromise even before GCS or other clinical measures have substantially deteriorated, and that not all sedative overdose presentations should be treated as physiologically equivalent.

The relationship between pCO₂ severity and adverse outcomes across hypercapnic patients has not been previously reported in a sedative‐specific cohort. The finding that pCO₂ ≥ 65 mmHg was associated with an intubation rate of 31.8% and ICU admission rate of 50.0%, compared with approximately 8.9% and 17.8% respectively for pCO2 of 45–54 mmHg, suggests that pCO₂ may offer a clinically useful threshold for escalation. A pCO₂ of 65 mmHg represents severe hypoventilation (roughly corresponding to an alveolar ventilation of approximately 50% of normal), and its presence in a conscious but obtunded patient warrants urgent airway assessment regardless of total GCS score [11]. This observation is hypothesis‐generating and requires prospective validation and also raises the question of whether non‐invasive ventilation strategies may have a role in patients with intermediate pCO₂ elevation prior to definitive airway intervention.

Lactate profiles differed by agent class, extending prior work demonstrating that benzodiazepine co‐ingestion was inversely associated with hyperlactatemia in a mixed overdose cohort [7]. Our data confirm this finding in a sedative‐specific cohort and further demonstrate that GHB also produces low lactate, consistent with its GABAergic mechanism and absence of sympathomimetic or metabolic toxicity. In contrast, ethanol produced elevated lactate in nearly half of presentations, reflecting its well‐described inhibition of gluconeogenesis and contribution to Type B lactic acidosis [6]. The elevated lactate associated with anti‐depressants (33%) may reflect toxidrome‐specific mechanisms including serotonin syndrome, seizure activity, or direct mitochondrial effects of tricyclic agents. These findings suggest that lactate interpretation should be contextualised by the likely causative agent class.

Among patients with venous respiratory acidosis, the co‐occurrence of elevated lactate (n = 15) was associated with an ICU admission rate of 53.3% compared with 36.8% in those with venous respiratory acidosis and normal lactate. This observation is limited by small numbers and should be considered descriptive only. The co‐occurrence of hypercapnic respiratory failure and elevated lactate may identify patients with concurrent ventilatory compromise and metabolic derangement and warrants evaluation in larger prospective cohorts.

The absence of an association between number of agents ingested and acid–base phenotype classification suggests that agent pharmacology rather than agent number is the primary determinant of acid–base disturbance in sedative toxidrome. An interaction between polypharmacy and outcomes within the venous respiratory acidosis subgroup was observed but was limited to 10 patients with polypharmacy, precluding meaningful inference; this relationship warrants evaluation in larger prospective datasets.

Venous pH outperformed all individual VBG components in discriminating outcomes (AUC 0.631–0.683 vs. 0.546–0.600 for other parameters), consistent with the approach of our primary analysis which used pH as a continuous predictor [3]. pH integrates both the respiratory and metabolic axes and provides a single, clinically actionable value [12]. The characterisation of the underlying pattern—whether hypercapnia is accompanied by frank venous acidaemia or occurs with a preserved pH—adds mechanistic context that continuous pH measurement alone cannot provide.

The absence of VBG in 44 patients (11.8%), all of whom had 0% intubation and 0% ICU admission, is consistent with clinical assessment guiding selective VBG use in lower acuity presentations. This pattern also suggests that the VBG data in this cohort have high clinical relevance and are not subject to significant ascertainment bias toward low‐acuity cases.

4.1. Limitations

Several limitations warrant acknowledgement. The pure metabolic and mixed acidosis groups were very small (n = 3 and n = 4, respectively), precluding reliable statistical analysis and limiting conclusions in these categories to descriptive observation. It is acknowledged that intubation decisions in clinical practice may be influenced by VBG findings, including pCO₂ values; this introduces the possibility that the associations observed between acid–base parameters and intubation partly reflect clinical decision thresholds rather than purely physiological risk. Prospective studies using pre‐specified escalation criteria independent of VBG values would be required to address this limitation. Anion gap was not calculable from the available dataset as sodium and chloride were not consistently extracted, preventing mechanistic sub‐classification of metabolic acidosis. Serial VBG measurements were not analysed; the dynamic evolution of acid–base status over time could not be assessed from this cross‐sectional dataset. Agent class was assigned based on clinical documentation at the time of the toxicology consultation; in polypharmacy presentations, misclassification of the primary agent cannot be excluded. The interaction between polypharmacy and outcomes within the venous respiratory acidosis subgroup was observed in only 10 patients with polypharmacy and should be considered underpowered for inference. Finally, although this is a consecutive cohort from two sites within a single health district, generalisability to settings with different drug supply profiles or clinical practices may be limited. Prospective external validation of the pCO₂ severity threshold and the combined respiratory acidosis‐plus‐elevated‐lactate finding is warranted.

5. Conclusions

When venous‐specific reference ranges are applied, venous respiratory acidosis occurred in 16.4% of sedative overdose patients with VBG data and was associated with ICU admission in 40.7% and complications in 64.8% of cases. Hypercapnia burden varied by agent class, being most common in opioid exposures and least common in anti‐psychotic presentations, consistent with pharmacological differences in respiratory centre depression. Among hypercapnic patients, a pCO₂ ≥ 65 mmHg identified a subgroup with ICU admission approaching 50%. Venous pH retains the strongest individual discriminative performance for outcomes. These findings support routine early VBG measurement in sedative overdose presentations and highlight the importance of applying venous‐specific reference ranges when interpreting point‐of‐care acid–base results in this context.

Funding

The authors have nothing to report.

Ethics Statement

This study was approved by the Western Sydney Local Health District Human Research Ethics and Governance Office as a secondary analysis of previously approved data.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Table S1: In‐hospital complications by acid–base phenotype.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.