First-attempt success and associated factors among emergency tracheal intubations in two addis Ababa hospitals Authors: Tsion K. Admas, Biruk T. Mengistie, Chernet T. Mengistie, Mikiyas G. Teferi, Meron H. Biza, Tseganesh G. Kebede, Getaw W. Hassen, Merahi K. Merahi Published: 2025 ABSTRACT IntroductionFirst-pass success (FPS) in emergency tracheal intubation is a key quality metric linked to fewer intubation-related complications, but data from low-income settings are limited. This study aimed to measure FPS and identify intubation-related complications in two tertiary hospitals in Addis Ababa, Ethiopia.MethodsWe conducted a prospective, cross-sectional observational study of consecutive emergency and ICU tracheal intubations at Tikur Anbessa Specialized Hospital (TASH) and Zewditu Memorial Hospital (ZMH) from May to October 2024. Patients aged 13 years or older undergoing emergency intubations were included. Online tool was developed to collect key variables from clinicians. The primary outcome was FPS (successful endotracheal tube placement on the first laryngoscope insertion). Bivariate analyses and multivariable logistic regression were used to identify independent predictors; adjusted odds ratios (aOR) with 95% confidence intervals (CI) were reported.ResultsA total of 112 intubations were analyzed. Median age of patients was 37.5 years (IQR 25–55); 62.5% were male. Hypoxic respiratory failure was the predominant indication (68.8%). Direct laryngoscopy was used in all cases; the most common induction strategy was ketamine (50/112 [44.6%]) with succinylcholine (67/112 [59.8%]) as the neuromuscular blocking agent. FPS was 64.3% (72/112). Additional successes occurred on the second (25.0%), third (8.0%), and ≥ 4 attempts (2.7%). One or more complications occurred in 42 out of the 112 cases (37.5%; 95% CI 28.5–46.5%), highlighting a substantial immediate adverse-event burden; cardiovascular instability occurred in 22.3%, cardiac arrest in 5.4%, and death within one hour in 2.7%. On adjusted analysis, operator training level and airway visualization were the strongest predictors: Year-II residents (aOR 46.81; 95% CI 3.03–722.72; p = 0.006) and Year-III residents (aOR 406.30; 95% CI 13.72–12,033.01; p < 0.001) had markedly higher odds of FPS compared to Year-I residents. Intubations without anticipated difficulty were more likely to succeed (aOR 10.74; 95% CI 1.01–114.45; p = 0.049). A Cormack–Lehane grade III view predicted failure (aOR 0.005; 95% CI 0.000–0.828; p = 0.042), while an abducted vocal cord favored success (aOR 23.96; 95% CI 2.93–195.76; p = 0.003).ConclusionFPS in these Addis Ababa hospitals (Tikur Anbessa Specialized Hospital and Zewditu Memorial Hospital) was 64.3%, a level we consider suboptimal compared with commonly cited benchmarks of ≥ 80%. This is lower than pooled benchmarks reported from large high-income country series (~ 79–84%) but within the range reported in some low- and middle-income settings. Operator experience and airway visualization were the dominant determinants of one-pass success. Given the high immediate complication rate (37.5%), targeted supervised training, anticipation of difficult anatomy, and prioritized expansion of airway adjuncts (e.g., bougie, phased introduction of video laryngoscopy and capnography) should be implemented to improve FPS and reduce harm in resource-limited emergency settings. Introduction Emergency tracheal intubation is a high-acuity procedure performed worldwide in both emergency departments and intensive care units for critically ill patients who require definitive airway or ventilatory support [1, 2]. Common indications include respiratory failure, airway protection (e.g. decreased level of consciousness), and trauma [2, 3]. Emergency tracheal intubations are often performed in patients with severe physiologic derangements and under time pressure; these factors, patient instability and limited time for preparation, increase the risk of hypoxemia, hypotension, and cardiac arrest during the procedure [1, 4]. Contemporary guidelines therefore, emphasize both pre-intubation preparation and intra-procedural strategies to reduce per-intubation adverse events. For example, structured checklists, careful preoxygenation, hemodynamic optimization and limiting the number/duration of laryngoscope insertions [5, 6]. First-pass success (FPS) is defined as successful endotracheal tube placement on the first laryngoscope insertion [1]. It is a widely accepted quality metric [1, 7]. Higher FPS rates are associated with fewer airway-related complications, and complications arise increasingly with each additional attempt [8]. Pooled analyses and large emergency department (ED) series from high-income settings typically report FPS in the ~ 80–85% range (Park et al. meta-analysis pooled FPS ≈ 84%); rates < 80% are generally considered suboptimal. these figures are frequently used as benchmarks for technical performance in hospital EDs, while higher prehospital or selected high-performing system rates (> 90%) reflect different settings and referral/skill mixes and are therefore less directly comparable [7]. Each extra attempt substantially raises the odds of hypoxia and hemodynamic instability [2, 7, 9]. Because major peri-intubation adverse events are common in critically ill patients (for example, a large international ICU study reported nearly half of patients experienced at least one major adverse event during intubation) [8], maximizing FPS is critical to reduce the duration of hypoxia and the risk of airway-related complications such as aspiration and cardiovascular instability [1, 8, 10]. Various interventions and predictors have been studied to optimize FPS [10–12]. Rapid-sequence intubation (RSI) with sedative induction and neuromuscular blockade is recommended for patients who are rapidly deteriorating [11, 13]. Observational data show that RSI significantly improves FPS compared to sedation-only techniques [11]. Proper preoxygenation, pre-intubation checklists, and hemodynamic optimization are also advocated [14, 15]. Additionally, Video laryngoscopy and the use of an intubating bougie as a first-line adjunct have similarly shown benefits [10, 16–18]. Patient and airway features (limited mouth opening, short neck, obesity, presence of blood/secretions) predict difficult laryngoscopy and therefore lower FPS; recognizing these predictors allows clinicians to pre-emptively select adjuncts or the most experienced operator [7, 13]. Cormack–Lehane grade III/IV views drastically lower FPS [19, 20]. Conversely, visualization of open vocal cords facilitates tube passage [7, 9, 19]. Clinician experience is a major factor: less-experienced providers typically have lower FPS [5, 18]. Data from NEAR and other registries confirm that trainee year and specialty are independent predictors of success [18]. Cardiopulmonary status also matters; Physiologic derangements (severe hypoxemia, hypotension, acidosis) create a ‘physiologically difficult’ airway, not by altering anatomy, but by shortening the time available for safe attempts and increasing the risk of immediate complications, and thus can reduce FPS and worsen outcomes [2, 14]. In resource-limited settings, the clinical consequences of failed or repeated attempts are amplified because advanced airway equipment, continuous capnography, and immediate critical-care backup may be unavailable [21, 22]; therefore, improving FPS has the potential not only to reduce individual patient harm but also to ease pressures on scarce ICU resources [22]. But, few studies report FPS in such settings [3, 21, 22]. In this context, we evaluated FPS and its predictors among intubations at two tertiary hospitals in Ethiopia, using the above definitions and examining interventions known to enhance success. Methods Study design and setting We conducted a prospective, cross-sectional observational study of emergency tracheal intubations performed at two state-funded hospitals in Addis Ababa, Ethiopia: Tikur Anbessa Specialized Hospital (TASH) and Zewditu Memorial Hospital (ZMH). TASH is a tertiary referral (a major specialist center receiving referrals from across the region) and teaching hospital that provides care for approximately 370,000–400,000 patients per year and manages around 18,000 emergency department visits annually. The TASH ED resuscitation (“red”) area, the high-acuity resuscitation bay used for critically ill patients, has seven beds and two mechanical ventilators. ZMH is a university-affiliated teaching hospital with an intensive care unit (10 beds, eight functional mechanical ventilators); unlike TASH it functions primarily as a regional/teaching hospital rather than the city’s main tertiary referral center. The study period was 6 months (May 2024 to October 2024). We selected a prospective design because an initial audit of routine intubation charts showed frequent absence or inconsistent recording of key, time-sensitive variables required for accurate ascertainment of first-pass success (e.g., Cormack–Lehane grade, number of laryngoscope insertions, tube-insertion attempts, device/adjunct used). A brief structured form was therefore completed immediately after each procedure to ensure complete capture of these variables. Participants and eligibility All patients aged ≥ 13 years who underwent emergency or ICU tracheal intubation at TASH ED or ZMH ED/ICU during the study period were eligible. We excluded patients who had already been intubated elsewhere upon arrival. Intubations performed by emergency medicine and anesthesia residents and consultants were included. All eligible cases meeting the predefined inclusion criteria were enrolled during the study period. Sample size and sampling procedure The sample size for the primary objective (first-pass success proportion) was calculated using a single-proportion formula with the following parameters: Z = 1.96 (95% confidence level), assumed proportion p = 0.706 (70.6% from a prior Ethiopian study), and margin of error d = 0.03 (3%). This yields an initial sample size n₀ ≈ 886. Because the estimated 6-month (May–October 2024) sampling frame from the two hospitals was small (N ≈ 120), a finite-population correction was applied: n = n₀ / [1 + (n₀ − 1)/N] ≈ 106. A 5% nonresponse adjustment was then added, giving a final required sample size of 112. Secondary-objective sample size requirements exceeded the practical sampling frame; therefore the primary-objective–based sample (n = 112) was used. The calculation was cross-checked using Epi Info. Data collection instrument and procedure We adapted a structured questionnaire from prior studies (INTUBE/large airway studies and local Ethiopian work) and implemented it as an English-language Google Form [2, 3]. Subject-matter experts and senior emergency/anesthesia clinicians reviewed the content for validity, and the instrument was pilot-tested on 5% of the calculated sample in the study area with subsequent revision. The structured form captured patient demographics, pre-intubation status, peri-intubation variables (including indications, oxygenation/positioning, devices, and medications), and immediate post-intubation outcomes and complications. Participating physicians received brief training on form completion and the form was intentionally concise. The structured form supplemented routine clinical documentation but did not replace it. To minimise extra burden we pilot-tested the instrument, kept fields short and mostly mandatory, and required completion only after active resuscitation and once the patient was stabilized. Completion was therefore an additional but brief post-procedure documentation step rather than an in-the-moment intervention. All clinicians who performed eligible emergency or ICU intubations at the study sites during the study period were invited to participate. Completed forms were submitted to a secure, password-protected study account (institutional Google Workspace account restricted to the study team and protected by multi-factor authentication). Written informed consent was obtained from each participating physician and the institutional review board approved the study. All data were collected prospectively by the airway operator immediately after the procedure using the structured form. Several items on the form were operator-reported rather than independently instrument-measured. Specifically, appropriate positioning and lifting force were recorded as the operator’s judgment (no goniometer or force-meter was available); aspiration was recorded only if it was directly witnessed or suctioned during the procedure (no routine radiographic confirmation). Definitions and study variables The primary outcome was first-pass success (FPS), defined as correct tracheal placement of the endotracheal tube on the first laryngoscope insertion and first tube insertion attempt (i.e., one laryngoscope insertion and one ETT pass). Secondary outcomes were intubation-related complications recorded as events occurring during laryngoscopy/attempts or within the immediate post-procedure window. These included: esophageal intubation; hypoxemia during attempts (SpO₂ ≤80% during any attempt); cardiovascular instability (sustained SBP < 90 mmHg or the initiation/escalation of vasopressor support within 10 min of the procedure); cardiac arrest; death within 1 h; and other airway injuries (e.g., mucosal trauma, aspiration, pneumothorax). Plan A was defined as the primary intended technique (e.g., direct laryngoscopy with stylet or bougie-assisted intubation); Plan B as the immediate backup manoeuvre (e.g., change of blade or operator, or insertion of a supraglottic airway for temporary ventilation); and Plan C as the rescue/emergency front-of-neck airway (e.g., cricothyroidotomy). Cardiovascular instability was defined as a systolic blood pressure below 90 mmHg or the initiation of a new vasopressor within 10 min following intubation. The study form included options to record ETT confirmation by waveform or colorimetric capnography and by clinical methods (direct visualization of tube passage through the vocal cords, chest rise, and auscultation). Current airway guidelines consider continuous waveform capnography the gold standard for confirming tracheal placement and reducing unrecognised oesophageal intubation [6]. However, during the study period, placements were confirmed mainly by direct laryngoscopic visualization and auscultation (see Results), as routine waveform capnography was not available at the study sites. These choices likely introduce measurement and classification bias for the affected variables; we discuss the expected direction and implications of these biases in the Limitations section. Operational definitions for key subjective variables are provided as the following. Pre-oxygenation was recorded by device (non-rebreather/reservoir mask, bag-mask ventilation, CPAP/BiPAP, high-flow nasal cannula, none). Apneic oxygenation was defined as delivery of supplemental oxygen during laryngoscopy/apnoea (for example, low-flow nasal cannula or high-flow nasal oxygen continuing during attempts). Appropriate patient positioning was recorded as a binary, operator-judged item completed immediately after the procedure. Lifting force was recorded as an operator-judged binary variable (Yes/No) indicating whether the intubator perceived that increased upward lifting force was required during laryngoscopy; this was not objectively quantified with a measurement device. Body habitus was recorded by the intubating clinician as an operator-estimated BMI and categorized as lean (BMI < 30 kg/m²) or obese (BMI ≥ 30 kg/m²). When both height and weight were available in the medical record, a calculated BMI was used in preference to the operator estimate. Aspiration was recorded when regurgitation of gastric contents into the oropharynx or airway was directly witnessed or when new oropharyngeal/tracheal material was suctioned during or immediately after the procedure. This clinically observed definition is commonly used in observational airway studies where routine radiographic or bronchoscopic confirmation is not feasible; however, it likely underestimates the true incidence compared with systematic radiographic or bronchoscopic screening [2, 9] (see Limitations). Injury to the oropharynx was recorded when the intubating clinician or nursing staff documented visible mucosal trauma, bleeding, or laceration during the procedure or on immediate post-intubation inspection/suctioning; specialist (ENT/bronchoscopic) assessments were included when available. ET tube misplacement was defined as incorrect initial placement of the endotracheal tube into the oesophagus or a mainstem bronchus identified at the time of the attempt; it did not include post-procedure dislodgement. Airway protection was defined as intubation performed because the treating clinician judged that the patient was unable to reliably protect their airway. Data quality and management To reduce reporting bias and improve completeness, the principal investigator reviewed submissions daily (within 24 h), cross-checked entries with nursing documentation and medical records during routine checks, and queried operators promptly when discrepancies were found. No clinical airway training or changes to therapeutic management were introduced as part of the study; the only modification to routine practice was brief, post-procedure documentation using the standardized form. Pilot testing and brief training minimised additional clinician burden, forms were completed after the acute procedure to avoid interfering with care, and the IRB judged the incremental documentation burden acceptable relative to the anticipated benefit of locally relevant safety data. Each Google Form entry required mandatory responses for most items to reduce missing data. Entries were coded and exported to Microsoft Excel for cleaning and then imported to SPSS v26 for analysis. Patient medical record numbers were recorded for de-duplication and verification; operator identifiers were not collected to preserve anonymity. Statistical analysis Data were checked for completeness and outliers before analysis. Continuous variables were summarized using medians and interquartile ranges (IQR) because distributions were non-normal; categorical variables were presented as counts and percentages. Bivariate analyses employed chi-square tests for categorical variables and the Mann-Whitney U test for comparisons involving continuous variables. Variables with p < 0.25 on bivariate analysis were entered into multivariable binary logistic regression to identify independent factors associated with FPS. Model fit was assessed using the Hosmer–Lemeshow goodness-of-fit test; multicollinearity and influential outliers were evaluated and addressed if present. Adjusted odds ratios (aOR) with 95% confidence intervals (CI) were reported, and p < 0.05 was considered statistically significant. Analyses were performed using SPSS v26. Because the number of first-pass failures was limited, some estimates in the multivariable model produced wide confidence intervals. We reviewed variable distributions to ensure adequate cell counts and checked for collinearity before model fitting. The final model was retained because all included predictors had clear clinical relevance and acceptable data support. Ethical considerations Ethical approval for this study was obtained from the Institutional Review Board of of Addis Ababa University College of Health Sciences (IRB approval no. AAU-CHS/MS/2024/201). Additional permissions were secured from the Addis Ababa Health Bureau. All participants provided written informed consent after being informed of the study’s aims, risks, and benefits. Confidentiality was maintained by assigning numeric identifiers and securely storing data. All procedures performed in this study involving human participants were conducted in accordance with the ethical standards of the institutional and national research committees and with the 1964 Declaration of Helsinki and its later amendments. To minimise burden on providers and avoid interference with patient care, the data-collection form was intentionally brief and completed only after the procedure, not during active resuscitation. The IRB reviewed the procedures and judged that the brief, post-procedure documentation (pilot-tested for brevity), together with the potential benefits of generating locally relevant safety and quality data, made the minimal additional workload ethically acceptable. Results Demographics and characteristics of patients undergoing intubations A total of 112 emergency tracheal intubations performed during the study period were included. The cohort had a median age of 37.5 years (IQR 25–55); 62.5% were male. More than half (56.2%) had at least one comorbidity, with hypertension the most frequent comorbidity. Non-traumatic decreased level of consciousness (25.9%), sepsis/septic shock (19.6%), and Acute Respiratory Distress Syndrome (ARDS) (17%) were the most typical presenting conditions. Before intubation nearly half of the patients (47.3%) were supported with a non-rebreather mask (Table 1). Table 1Baseline characteristics of intubated patients, including demographics, comorbidities, clinical conditions, pre-intubation vital signs, and oxygen supportPatient characteristicsSubclassn (%)Age median (IQR)_Yr37.5 (25–55)SexMale70 (62.5)Female42 (37.5)Body habitusLean88 (78.6)Obese24 (21.4)ComorbidityYes63 (56.2)No49 (43.8)Types of Comorbidities(n = 63)Hypertension27 (42.9)Diabetes18 (28.6)Malignancy (hematologic or solid)8 (12.7)Asthma/ Chronic obstructive pulmonary disease (COPD)10 (15.8)Others28 (44.3)Presenting condition/complaintSepsis or septic shock22 (19.6)decreased level of consciousness (non-traumatic)29 (25.9)Traumatic brain injury (TBI)18 (16.1)ARDS19 (17.0)Poisoning4 (3.6)Acute abdomen5 (4.5)Guillain–Barré syndrome (GBS)3 (2.7)Others12 (10.8)Pre-intubation vital signsMedian (IQR)Systolic blood pressure (SBP)110 (90–130)Pulse rate (PR)118 (103–129)Respiratory rate (RR)40 (30–48)Oxygen saturation (SPO2)86 (76–90)Preoxygenation methodNasal prong10 (8.9)Simple face mask34 (30.4)Non-rebreather mask53 (47.3)CPAP/ BiPAP15 (13.4)Specific comorbidity counts and percentages are reported among patients with any comorbidity (n = 63); percentages in that subsection are therefore relative to 63. Percentages for all other categorical variables are relative to N = 112.IOR = inter-quartile range; COPD = chronic obstructive pulmonary disease; SBP = systolic blood pressure; CPAP/ BiPAP = continuous positive airway pressure / bilevel positive airway pressure Table 1Baseline characteristics of intubated patients, including demographics, comorbidities, clinical conditions, pre-intubation vital signs, and oxygen supportPatient characteristicsSubclassn (%)Age median (IQR)_Yr37.5 (25–55)SexMale70 (62.5)Female42 (37.5)Body habitusLean88 (78.6)Obese24 (21.4)ComorbidityYes63 (56.2)No49 (43.8)Types of Comorbidities(n = 63)Hypertension27 (42.9)Diabetes18 (28.6)Malignancy (hematologic or solid)8 (12.7)Asthma/ Chronic obstructive pulmonary disease (COPD)10 (15.8)Others28 (44.3)Presenting condition/complaintSepsis or septic shock22 (19.6)decreased level of consciousness (non-traumatic)29 (25.9)Traumatic brain injury (TBI)18 (16.1)ARDS19 (17.0)Poisoning4 (3.6)Acute abdomen5 (4.5)Guillain–Barré syndrome (GBS)3 (2.7)Others12 (10.8)Pre-intubation vital signsMedian (IQR)Systolic blood pressure (SBP)110 (90–130)Pulse rate (PR)118 (103–129)Respiratory rate (RR)40 (30–48)Oxygen saturation (SPO2)86 (76–90)Preoxygenation methodNasal prong10 (8.9)Simple face mask34 (30.4)Non-rebreather mask53 (47.3)CPAP/ BiPAP15 (13.4)Specific comorbidity counts and percentages are reported among patients with any comorbidity (n = 63); percentages in that subsection are therefore relative to 63. Percentages for all other categorical variables are relative to N = 112.IOR = inter-quartile range; COPD = chronic obstructive pulmonary disease; SBP = systolic blood pressure; CPAP/ BiPAP = continuous positive airway pressure / bilevel positive airway pressure Baseline characteristics of intubated patients, including demographics, comorbidities, clinical conditions, pre-intubation vital signs, and oxygen support Types of Comorbidities (n = 63) Pre-intubation vital signs Median (IQR) Specific comorbidity counts and percentages are reported among patients with any comorbidity (n = 63); percentages in that subsection are therefore relative to 63. Percentages for all other categorical variables are relative to N = 112. IOR = inter-quartile range; COPD = chronic obstructive pulmonary disease; SBP = systolic blood pressure; CPAP/ BiPAP = continuous positive airway pressure / bilevel positive airway pressure Most procedures took place in the ZMH ICU (42.9%, n = 48 ), with the remainder performed in the TASH and ZMH emergency departments. A single physician performed the intubation (i.e., one clinician actively executing the tube passage) in 61.6% (n = 69) of cases; in the remainder two or more team members were directly involved. Year-II and Year-III Emergency medicine and critical care (EMCC) residents performed the majority of procedures (46.4%, n = 52 and 42.9%, n = 48 respectively). Type 1 (Hypoxic) respiratory failure was the predominant indication for intubation (68.8%, n = 77); other common indications included airway protection (27.7%, n = 31) and Type 2 (Hypercapnic) respiratory failure (15.2%, n = 17). 86.6% (n = 97) of patients were being intubated for the first time during the current admission (Table 2). Table 2Distribution of intubation characteristics (place, number of operators, operator specialty/experience, indication, and single versus multiple intubation attempts)Subclassn (%)Place of intubationTASH Emergency33 (29.5)ZMH Emergency31 (27.7)ZMH ICU48 (42.9)Number of operators involved per each intubation1 physician69 (61.6)2 physicians30 (26.8)3 physicians13 (11.6)Operator’s specialty and experienceYear I EMCC4 (3.6)Year II EMCC52 (46.4)Year III EMCC48 (42.9)EMCC Consultant4 (3.6)Anesthesia4 (3.6)Indication for intubationAirway protection31 (27.7)Type 1 (Hypoxic) respiratory failure77 (68.8)Type 2 (Hypercapnic) respiratory failure17 (15.2)Anticipated clinical deterioration5 (4.7)Single versus multiple intubation attemptsFirst intubation97 (86.6)Re- intubation15 (13.4)Footnote: Type 1 (Hypoxic) respiratory failure: primary oxygenation failure (low PaO₂ or SpO₂ despite supplemental oxygen). Type 2 (Hypercapnic) respiratory failure: primary ventilatory failure (elevated PaCO₂ or clinical evidence of ventilatory failure)TASH = Tikur Anbessa Specialized Hospital; ZMH = Zewditu Memorial Hospital; EMCC = Emergency medicine and critical care Table 2Distribution of intubation characteristics (place, number of operators, operator specialty/experience, indication, and single versus multiple intubation attempts)Subclassn (%)Place of intubationTASH Emergency33 (29.5)ZMH Emergency31 (27.7)ZMH ICU48 (42.9)Number of operators involved per each intubation1 physician69 (61.6)2 physicians30 (26.8)3 physicians13 (11.6)Operator’s specialty and experienceYear I EMCC4 (3.6)Year II EMCC52 (46.4)Year III EMCC48 (42.9)EMCC Consultant4 (3.6)Anesthesia4 (3.6)Indication for intubationAirway protection31 (27.7)Type 1 (Hypoxic) respiratory failure77 (68.8)Type 2 (Hypercapnic) respiratory failure17 (15.2)Anticipated clinical deterioration5 (4.7)Single versus multiple intubation attemptsFirst intubation97 (86.6)Re- intubation15 (13.4)Footnote: Type 1 (Hypoxic) respiratory failure: primary oxygenation failure (low PaO₂ or SpO₂ despite supplemental oxygen). Type 2 (Hypercapnic) respiratory failure: primary ventilatory failure (elevated PaCO₂ or clinical evidence of ventilatory failure)TASH = Tikur Anbessa Specialized Hospital; ZMH = Zewditu Memorial Hospital; EMCC = Emergency medicine and critical care Distribution of intubation characteristics (place, number of operators, operator specialty/experience, indication, and single versus multiple intubation attempts) Footnote: Type 1 (Hypoxic) respiratory failure: primary oxygenation failure (low PaO₂ or SpO₂ despite supplemental oxygen). Type 2 (Hypercapnic) respiratory failure: primary ventilatory failure (elevated PaCO₂ or clinical evidence of ventilatory failure) TASH = Tikur Anbessa Specialized Hospital; ZMH = Zewditu Memorial Hospital; EMCC = Emergency medicine and critical care Pre-intubation preparation, difficulty anticipation, and optimization measures are shown in Table 3. Appropriate patient positioning, team role assignment, equipment/back-up preparation, and verbalized contingency plans were common. Physiologic difficulty was noted in the majority of cases where difficulty was recorded, and hemodynamic optimization (isotonic saline ± vasopressors) was performed when indicated. Rapid sequence intubation (RSI) and bag-mask ventilation (BMV) were frequently used for preoxygenation. Ketamine was the most common induction agent and direct laryngoscopy was used in all intubations. Table 3Comparison of intubation preparation, difficulty anticipation, optimization, sequence, pre-oxygenation, and medications between first-pass success and unsuccessful intubationsSubclassFirst-pass successfulNo/ percentFirst-pass unsuccessfulNo/ percentIntubation preparation and planning(n = 112)(Pre-procedural checklist)Appropriate patient positioning59/ 81.9%28/ 70%Assessment of potential airway difficulty45/ 62.5%18/ 45%Verbalized ‘’Plan A’’, ‘’Plan B’’, ‘’Plan C’’32/ 44.4%17/ 42.5%Equipment and backup preparation44/ 61.1%25/ 62.5%Assigned individual roles for team members53/ 73.6%27/ 67.5%Post-intubation sedation and analgesia preparation46/ 63.9%22/ 55%Difficulty anticipation(n = 90)Difficult laryngoscopy19/ 35.8%13/ 35.1%Difficult BMV11/ 20.8%10/ 27%Difficult Extra glottic device0/ 0.0%3/ 8.1%Difficult cricothyroidotomy0/ 0.0%1/ 2.7%Physiologic difficulty38/ 71.1%24/ 64.9%Hemodynamic optimization (n = 68)Isotonic saline 20 ml/kg42/ 97.7%23/ 92%Vasopressors19/ 44.2%9/ 36%Blood products1/ 2.3%1/ 4%Intubation modalitySedation only intubation21/ 29.2%7/ 17.5%Rapid sequence intubation (RSI)37/ 51.4%30/ 75%Delayed sequence intubation (DSI)1/ 1.4%0/ 0.0%Crush intubation13/ 18.1%3/ 7.5%Pre-oxygenation techniqueNon-rebreather mask11/ 15.3%5/ 12.5%BMV52/ 72.2%29/ 75.2%CPAP/ BiPAP2/ 2.8%2/ 5%High-flow nasal cannula (HFNC)5/ 6.9%3/ 7.5%None2/ 2.8%1/ 2.5%Neuromuscular blocking agentNone9/ 22.5%34/ 47.2%Succinylcholine31/ 77.5%36/50%Rocuronium0/ 0.0%1/ 1.4%Vecuronium0/ 0.0%1/ 1.4%Induction agentNo induction agent3/7.5%13/ 18.1%Ketamine16/ 40%34/ 47.2%Propofol11/ 27.5%15/ 20.8%Ketamine and propofol10/ 25%10/ 13.9% Table 3Comparison of intubation preparation, difficulty anticipation, optimization, sequence, pre-oxygenation, and medications between first-pass success and unsuccessful intubationsSubclassFirst-pass successfulNo/ percentFirst-pass unsuccessfulNo/ percentIntubation preparation and planning(n = 112)(Pre-procedural checklist)Appropriate patient positioning59/ 81.9%28/ 70%Assessment of potential airway difficulty45/ 62.5%18/ 45%Verbalized ‘’Plan A’’, ‘’Plan B’’, ‘’Plan C’’32/ 44.4%17/ 42.5%Equipment and backup preparation44/ 61.1%25/ 62.5%Assigned individual roles for team members53/ 73.6%27/ 67.5%Post-intubation sedation and analgesia preparation46/ 63.9%22/ 55%Difficulty anticipation(n = 90)Difficult laryngoscopy19/ 35.8%13/ 35.1%Difficult BMV11/ 20.8%10/ 27%Difficult Extra glottic device0/ 0.0%3/ 8.1%Difficult cricothyroidotomy0/ 0.0%1/ 2.7%Physiologic difficulty38/ 71.1%24/ 64.9%Hemodynamic optimization (n = 68)Isotonic saline 20 ml/kg42/ 97.7%23/ 92%Vasopressors19/ 44.2%9/ 36%Blood products1/ 2.3%1/ 4%Intubation modalitySedation only intubation21/ 29.2%7/ 17.5%Rapid sequence intubation (RSI)37/ 51.4%30/ 75%Delayed sequence intubation (DSI)1/ 1.4%0/ 0.0%Crush intubation13/ 18.1%3/ 7.5%Pre-oxygenation techniqueNon-rebreather mask11/ 15.3%5/ 12.5%BMV52/ 72.2%29/ 75.2%CPAP/ BiPAP2/ 2.8%2/ 5%High-flow nasal cannula (HFNC)5/ 6.9%3/ 7.5%None2/ 2.8%1/ 2.5%Neuromuscular blocking agentNone9/ 22.5%34/ 47.2%Succinylcholine31/ 77.5%36/50%Rocuronium0/ 0.0%1/ 1.4%Vecuronium0/ 0.0%1/ 1.4%Induction agentNo induction agent3/7.5%13/ 18.1%Ketamine16/ 40%34/ 47.2%Propofol11/ 27.5%15/ 20.8%Ketamine and propofol10/ 25%10/ 13.9% Comparison of intubation preparation, difficulty anticipation, optimization, sequence, pre-oxygenation, and medications between first-pass success and unsuccessful intubations Intubation preparation and planning (n = 112) (Pre-procedural checklist) Difficulty anticipation (n = 90) Outcome and complications of tracheal intubations Immediately after intubation, median (Interquartile range) vital signs were: FiO₂ 100% (100–100), SpO₂ 96% (93–98), systolic blood pressure 100 mmHg (85–120), and pulse rate 115 beats per minute (BPM) (100–125.5). First-pass success (FPS) was achieved in 64.3% (n = 72) of intubations. An additional 25.0% (n = 28) were successful on the second attempt, 8.0% (n = 9) on the third, and 2.7% (n = 3) required more than three attempts. Endotracheal tube placement was most often confirmed by direct visualization (Fig. 1). Waveform capnography was not available at the study sites and was used in 0/112 (0%) of intubations. One or more complications occurred in 42 of 112 patients (37.5%; 95% CI 28.5–46.5%); cardiovascular instability was the leading complication category (22.3%). Cardiac arrest occurred in 5.4% and death within one hour in 2.7% of cases. Other complications were uncommon: pneumothorax (1.8%), vocal-cord injury (1.8%), aspiration (0.9%), oropharyngeal injury (0.9%), and failed airway (0.9%). Endotracheal tube misplacement was documented in 18.8% of patients; hypoxemia during attempts was uncommon (2.7%) (Table 4). Fig. 1Bar chart showing the outcome of tracheal intubations, with frequencies of successful attempts by number of trials and techniques used for ETT placement confirmation Bar chart showing the outcome of tracheal intubations, with frequencies of successful attempts by number of trials and techniques used for ETT placement confirmation Table 4Complications observed during and after tracheal intubationComplicationsSubclassn (%)Cardiovascular instability (Between induction and 1-hour post-intubation [n = 25])New or increased vasopressor administration11 (44)SBP less than 6510 (40)Arrhythmia1 (4)Cardiac arrest6 (24)Death3 (12)ET tube misplacement21 (18.8)Hypoxemia (SPO2 1 attempt raises odds of desaturation and hypotension several-fold [7–9]. Indeed, a 2024 ED registry found that the risk of major events increased from 1 attempt to 2 attempts (adjusted OR ~ 4.4) and to 3 attempts (OR ~ 13.9) [8]. Thus, our lower FPS likely contributed to the observed 37.5% event rate. We used only direct laryngoscopy (no video) and had many early trainees. In contrast, advanced airway adjuncts and experienced providers increase FPS [4, 12, 16]. For example, two large ED studies showed video laryngoscopy yielded significantly higher FPS (90.9%) than enhanced direct techniques (81.1%; AOR ≈ 2.8) [18], and trauma patients intubated by video had ~ 11% higher FPS than direct [17]. Similarly, bougie use has been shown to boost FPS from mid-80s to mid-90s [10, 16]. These interventions were underutilized or unavailable in our setting. Our multivariate analysis highlights that clinician skill and airway anatomy are paramount. The dramatic advantage of senior residents (aORs ~ 47 and 406) emphasizes the steep learning curve: inexperienced intubators often require more attempts [2, 4, 7]. This is consistent with literature: studies repeatedly find that providers with > 50 prior intubations or more senior training have higher FPS [4, 17]. We also found that anticipating difficulty predicted failure; when no difficulty was expected, FPS was > 10-fold higher. The Cormack-Lehane grade III view was essentially predictive of failure, a finding echoed in multiple studies that linked poor laryngeal view to failed intubation [17, 23]. In our data, both a favorable laryngeal view (lower Cormack–Lehane grade) and direct visualization of an abducted vocal cord were strongly associated with FPS. Although laryngeal view (graded by Cormack–Lehane) and vocal-cord position are related, they are distinct measures; visualization describes the overall glottic exposure while vocal-cord position describes the observed cord abduction, and both predicted success in our cohort [2, 7, 18]. Because first-pass failure events were relatively limited in number, our ability to precisely quantify the magnitude of some associations was constrained [24]. Clinically, this means that while the data consistently point to provider experience and equipment availability as important, modifiable factors, the exact effect sizes reported here should be viewed as exploratory rather than definitive [25]. The high complication rate observed (37.5%) and the consistent direction of associations across analyses nevertheless support prioritizing targeted training and equipment interventions [2]. Notably, after adjusting for these factors, neither pharmacologic interventions nor devices were found to be independently significant. This may reflect limited power or practice patterns. For example, although RSI was standard, it did not predict FPS in our model. This contrasts with Okubo’s multicenter study, where RSI doubled the odds of FPS [11]. It is possible that nearly all of our intubations were performed using RSI, which would reduce the detectable difference. Likewise, bougie use appeared higher in successes on bivariate analysis, but lacked an independent effect (Table 4). Given strong evidence that bougie improves success [10, 16], the fact that it was not significant here may result from low utilization or confounding by operator skill. In any case, our findings suggest that training and recognition of complicated anatomy should be priorities. These results have several implications. First, they highlight the gap between high-resource (HIC) EDs and low-resource (LMIC) settings such as ours. Interventions proven to improve FPS in the literature, such as routine bougie-first techniques [10, 16], use of videolaryngoscopes [17, 18, 23, 26], and standardized checklists [15], could be adapted where feasible, although uptake of these evidence-based practices in many LMIC settings has been slow due to resource and training constraints [22]. Second, the high complication rate reinforces the need for pre-oxygenation and preparation to mitigate hypoxemia/hypotension [2, 14]. Third, the importance of operator experience suggests that targeted training (including simulation or supervised practice) for junior clinicians could yield major improvements in one-pass intubations [27, 28]. Notably, even in extreme prehospital settings, high-performing systems emphasize FPS (e.g. FPS ~ 91% in Danish HEMS) and minimize attempts through experienced providers [29]; adopting similar standards should be a goal. Finally, we recommend a prioritized, feasible approach for LMIC settings: begin with supervised airway training, a short pre-intubation checklist, and universal access to a bougie; next, phase in portable capnography and pooled video laryngoscopes as resources allow. Conclusion In two Addis Ababa hospitals, first-pass success was suboptimal (64.3%), and immediate intubation-related complications were frequent (37.5%), indicating an urgent need for prioritized improvements. Success depended almost entirely on operator training level and airway visualization: senior EM trainees and easy anatomy were far more likely to achieve a single-pass intubation. Importantly, our findings suggest that enhancing training, anticipating potential difficulties, and utilizing adjuncts such as bougie-stylets or video laryngoscopy (when available) could substantially increase FPS in resource-limited settings. Priority should be given to supervised airway training, a brief pre-intubation checklist, and universal access to a bougie; portable waveform capnography and pooled video laryngoscopes can then be phased in as resources allow. Limitations This study was conducted in two urban tertiary hospitals in Addis Ababa and the single-city, two-center design with a moderate sample size limits precision and generalizability beyond similar urban Ethiopian tertiary centers. Although near-real-time clinician reporting and daily verification were used, data relied primarily on self-report and may have resulted in underreporting of minor adverse events. Procedural variables recorded by operators (e.g., apneic oxygenation, lifting force, positioning, and BMI estimate) were subjective and may be misclassified; these items should therefore be interpreted as exploratory. Routine waveform end-tidal CO₂ monitoring or waveform capnograph was not universally available, so ETT confirmation relied largely on visualization and auscultation and may have been misclassified. The lack of video laryngoscopes and limited anaesthesia involvement further restricts generalizability. Additionally, some adjusted estimates had wide 95% confidence intervals due to relatively few first-pass failures and sparse predictor cells, which limits confidence in the precise effect sizes. We also lacked long-term outcomes beyond the immediate period. Finally, the aspiration rate in our study may be underestimated because only witnessed or directly observed events were recorded, and radiographic or delayed detections were not captured.