Efficacy of abdominal thrusts and back blows for patients with foreign body airway obstruction: MOCHI registry analysis

Abstract

Aim

This study aimed to evaluate the effectiveness of abdominal thrusts and back blows as initial bystander interventions compared to the absence of bystander intervention in patients with foreign body airway obstruction (FBAO).

Methods

This was a prospective, observational study conducted at 25 hospitals in Japan (April 2020–March 2023), including patients aged ≥ 18 years presenting emergency departments with FBAO. The primary outcome was a favorable neurological outcome at 30 days, defined as a Cerebral Performance Category 1 or 2. Secondary outcomes were 30-day survival and success of obstruction relief. We used propensity scores with inverse probability of treatment weighting (IPTW) to adjust for confounding. Logistic regression and Cox proportional hazards models were applied for outcome analysis.

Results

Among 407 patients, 24 received abdominal thrusts, 76 received back blows, and 175 received no bystander intervention. Median age was 81. After IPTW adjustment, favorable neurological outcomes were significantly more frequent in patient receiving abdominal thrusts (38% vs. 16%; difference 22%, 95% CI 14%–31%) and back blows (31% vs. 16%; difference 15%, 95% CI 8%–23%) compared to no intervention. Back blows were also associated with higher survival (adjusted HR 0.52, 95% CI, 0.35–0.78), while abdominal thrusts were not (adjusted HR 0.73, 95% CI, 0.40–1.35).

Conclusion

Abdominal thrusts and back blows were both associated with significantly favorable neurological outcomes compared to no intervention. Back blows were additionally associated with higher survival.

Introduction

As individuals age or suffer neurological disorders, their ability to swallow deteriorates, increasing their risk of choking on foreign objects lodged in the airway.¹ Particularly in Japan, which is currently experiencing the most advanced population aging in the world, 35% of the population was over 60 years old as of 2022.² Consequently, over 8,000 people die annually from foreign body airway obstruction (FBAO) in Japan, making it the second leading cause of accidental death.³ In the United States (US), FBAO is also the fourth leading cause of accidental death.⁴ Thus, many developed countries are confronting the issue of FBAO, and it is expected to become an increasingly significant public health problem.

FBAO can quickly lead to severe hypoxia, and the prognosis is extremely poor if it escalates to cardiac arrest. Immediate removal of the foreign body is crucial to prevent cardiac arrest. If airway obstruction lasts six minutes or more, neurological outcomes may deteriorate.⁵ Even in countries with well-developed emergency medical systems, it is challenging for emergency medical services (EMS) to arrive at the scene within five minutes.⁶ Therefore, bystander intervention plays an important role in removing foreign bodies before EMS arrive.

Bystander removal procedures include back blows, abdominal thrusts, finger sweep, chest thrusts/compressions, and the use of airway clearance devices.⁷ The International Liaison Committee on Resuscitation (ILCOR) recommends starting with back blows in conscious patients with FBAO and moving to abdominal thrusts in adults and children older than one year if back blows prove ineffective.⁸ The previous study using MOCHI registry demonstrated that bystander intervention is associated with favorable neurological outcomes⁹; however, no studies have compared abdominal thrusts or back blows to a control group, and the most effective bystander technique for FBAO removal has yet to be identified.¹⁰ Although chest compressions have been studied in comparison to no intervention,¹¹ the evidence remains limited for other techniques.

This study aimed to fill a critical evidence gap by comparing the effectiveness of abdominal thrusts and back blows with no bystander intervention—an approach not previously taken in real-world FBAO research.

Methods

Study design

The MOCHI (Multi-center Observational CHoking Investigation) was a prospective　multi-center, observational study conducted at 25 hospitals in Japan, including all patients presenting to emergency departments with FBAO.¹² The study was approved by the Institutional Review Board of Nippon Medical School Hospital (Approval No. B-2019–019) as well as the ethics committees of each participating hospital.

Setting

The registry included all patients who were transported to the emergency department with FBAO by EMS. Patients who arrived without EMS transport were not included. FBAO was defined as life-threatening mechanical airway obstruction caused by foreign bodies, including both food and non-food materials. Patients were identified and registered by attending emergency physicians at each participating hospital using predefined inclusion criteria. A total of 25 hospitals, including eight university hospitals across Japan, participated in the registry from April 2020 to March 2023.

Patients were included if they had received either abdominal thrusts or back blows as initial intervention by non-emergency professional bystanders or had not received any bystander interventions prior to medical evaluation. Patients under 18 years of age were excluded, as the effectiveness of bystander interventions may vary with body size.

In Japan, the general public is instructed according to the Japan Resuscitation Council guidelines, which adhere to ILCOR recommendations.¹³ For adults with FBAO who are unable to cough, the guidelines recommend administering back blows first, followed by abdominal thrusts if the back blows are ineffective. Emergency medical dispatchers provided phone instructions on foreign body removal procedures with minimal regional variation due to government recommendations. Bystanders were advised to attempt only back blows unless they had been trained in abdominal thrusts. Due to the highly stringent criteria for terminating resuscitation efforts in the field, it is extremely rare for EMS not to transport patients to the hospital.

Data collection

A panel of board-certified emergency medicine experts developed the data collection form using electronic data capture tools, building on previous research.¹⁴ The registry documented demographic variables and specific variables related to FBAO, such as the type and location of the foreign body, characteristics of bystanders, types of interventions performed by both bystanders and healthcare professionals, success or failure of the interventions and outcomes at 30 days. The location of the foreign body in the airway was determined by emergency physicians based on clinical evaluation and imaging studies where available. It was categorized using the MOCHI classification, which consists of three types.¹⁴ If multiple interventions were performed, all were documented sequentially up to the seventh attempt. Data on FBAO incidents, including the characteristics of foreign body and bystander interventions, were prospectively collected through structured interviews conducted by treating physicians in the emergency department. Information was obtained primarily from EMS personnel and supplemented by accounts from bystanders, family members, and facility staff when available. The methods for data collection and quality assurance have been detailed elsewhere.¹²

Outcome measures

The primary outcome was defined as a favorable neurologic outcome, assessed by Cerebral Performance Category scale with scores of 1 (good recovery) or 2 (moderate disability) at 30 days.¹⁵ Secondary outcomes included survival at 30 days, prehospital cardiac arrest, and successful removal of airway obstruction, which was defined as the restoration of effective spontaneous breathing without ongoing signs of airway compromise. Patients who remained hypoxic, continued to cough severely, or had suspected residual airway foreign bodies or aspiration-related injury were not considered to have experienced successful removal. Data on intervention-related complications were collected through computed tomography scans, post-mortem imaging studies, or autopsies.

Statistical analysis

Descriptive statistics were used to summarize patient characteristics. Continuous variables are presented as medians and interquartile ranges, and were analyzed using the Mann–Whitney U‐test. Categorical variables were analyzed using the χ² test or Fisher’s exact test, as appropriate.

To evaluate the effectiveness of abdominal thrusts and back blows as a first bystander intervention respectively, two comparisons were made: abdominal thrusts versus no bystander interventions, and back blows versus no bystander interventions. Separate propensity score models were built for each comparison, and inverse probability of treatment weighting (IPTW) was applied accordingly.

Multiple imputation using predictive mean matching was performed to account for missing values in covariates and generate five complete datasets. Case with missing outcome data were excluded from the analysis. All exposure variables were mandatory and complete. To address potential confounding, propensity scores were calculated using a multivariable logistic regression model. The following variables were included in the propensity score model: patient age, sex, eating function, incident location (public or private), witness age, witness sex, bystander type (healthcare worker or not), EMS response time, and MOCHI classification (Type 1 obstruction). The variables used in this analysis were selected based on those used in the primary MOCHI registry analysis,⁹ to ensure consistency. Variable definitions are presented in Supplementary Table S1.

IPTW was then applied based on the calculated propensity scores to balance covariates between groups.¹⁶ After adjustment, the association between bystander intervention type and favorable neurological outcomes was assessed using weighted logistic regression. Cox proportional hazards models were used to evaluate 30-day survival, also adjusted using IPTW based on the same covariates as the logistic regression model.

A p-value of less than 0.05 was considered statistically significant. All statistical analyses were performed using R software, version 4.0.4 (The R Foundation for Statistical Computing, Vienna, Austria).

Results

Out of 407 patients with recorded 30-day outcomes, 24 patients received abdominal thrusts, 76 patients received back blows as a first bystander intervention, and 175 patients did not receive any bystander intervention in this study (Fig. 1). The frequency of multiple interventions is shown in Supplementary Figure A.

Fig. 1.

Study participant selection.

The median age was 81 years, with an approximately equal distribution between male and female patients. Dementia was the most frequently reported medical history. Over half of the patients were not independent in activities of daily living, yet three-quarters were able to eat independently. The most common locations for incidents were at home (57%), followed by nursing homes (19%), and restaurants (10%). Rice (24%), meat (17%), and mochi (rice cake) (16%) were the leading causes of choking (Table 1).

Table 1.

Patient characteristics.

	No bystander interventions (n = 175)	Bystander intervention
		Abdominal thrusts (n = 24)	P value	Back blow	P value
				(n = 76)
Age (median [IQR])	81	81.5	0.45	82	0.77
	[74.0–87.0]	[72.3–84.0]		[74.0–88.0]
Sex female, n (%)	79 (45%)	12 (50%)	0.82	37 (49%)	0.7
Past medical history
Dementia	53 (30%)	8 (33%)	0.95	19 (25%)	0.49
CVD	47 (27%)	5 (21%)	0.81	13 (17%)	0.23
Diabetes	13 (7%)	1 (4%)	0.87	8 (11%)	0.57
Parkinson’s disease	12 (7%)	2 (8%)	1	6 (8%)	0.98
CAD	8 (5%)	3 (13%)	0.24	3 (4%)	0.69
Schizophrenia	5 (3%)	2 (9%)	0.44	2 (3%)	1
Choking	3 (2%)	0 (0%)	0.78	2 (3%)	0.54
ADL independent	73 (42%)	10 (42%)	1	34 (45%)	0.76
Eating function			0.049		0.81
Independent	132 (75%)	20 (83%)		59 (78%)
Requiring assistance	25 (14%)	1 (4%)		10 (13%)
Completely dependent	16 (9%)	1 (4%)		7 (9%)
Unknown	2 (1%)	2 (8%)
Location			0.002		0.56
Home	103 (59%)	6 (25%)		47 (62%)
Nursing home	31 (18%)	6 (25%)		15 (20%)
Restaurant	19 (11%)	3 (13%)		6 (8%)
Group home	14 (8%)	5 (21%)		4 (5%)
Day care center	3 (2%)	1 (4%)		3 (4%)
Park	2 (1%)	1 (4%)		0 (0%)
Clinic	2 (1%)	0 (0%)		0 (0%)
Hospital	0 (0%)	1 (4%)		0 (0%)
School	0 (0%)	1 (4%)		0 (0%)
Street	0 (0%)	0 (0%)		1 (1%)
Vehicle	1 (1%)	0 (0%)		0 (0%)
Witness	125 (71%)	22 (92%)	0.06	73 (96%)	<0.001
Witness age (median [IQR])	60 [50–75]	50 [47–60]	0.003	50 [45–70]	0.003
Witness sex female, n (%)	69 (39%)	11 (46%)	0.1	37 (49%)	<0.001
Witness relation			<0.001		<0.001
Family	95 (54%)	8 (33%)		49 (65%)
Healthcare worker	17 (10%)	14 (58%)		23 (30%)
Friend	5 (3%)	0 (0%)		0 (0%)
Others	8 (5%)	0 (0%)		1% (1%)
Foreign body
Rice	42 (24%)	5 (21%)	0.93	18 (24%)	1
Mochi	30 (17%)	4 (17%)	1	10 (13%)	0.55
Bread	9 (5%)	2 (8%)	0.87	5 (7%)	0.88
Meat	24 (14%)	3 (13%)	1	20 (26%)	0.03
Fish	7 (4%)	0 (0%)	0.68	1 (1%)	0.47
Sushi	4 (2%)	0 (0%)	1	3 (4%)	0.75

ADL, activity of daily life; CAD, coronary artery diseases; CVD, cerebrovascular diseases; IQR, interquartile ranges.

Data were missing for 56 patients (20%), the majority of which (n = 55) involved witness-related variables, such as age or relationship to the patient. All other variables had complete or near-complete data, with missingness below 1%.

In the unadjusted analysis, patients who received abdominal thrusts or back blows had significantly higher rates of favorable neurological outcomes compared to those who did not receive any bystander interventions (37% vs. 15%; difference 23%, 95% confidence interval [CI] 3% to 43% for abdominal thrusts: 32% vs. 15%; difference 17%, 95% CI 5 to 29% for back blows). After adjusting for covariates and improving covariate balance (Supplementary Fig.B.1, Fig.B,2), both abdominal thrusts and back blows were associated with significantly higher proportions of favorable neurological outcomes compared to no bystander interventions (38% vs. 15%; differences 22%, 95% CI 14% to 31% for abdominal thrusts: 31% vs. 16%; difference 15%, 95% CI 8% to 23% for back blows) (Table 2). Abdominal thrusts were associated with significantly higher odds of favorable neurological outcomes, with an unadjusted odds ratio (OR) of 3.44 (95% CI, 1.36 to 8.67) and an adjusted OR of 3.37 (95% CI, 1.07 to 10.6). Similarly, back blows were associated with significantly higher odds of favorable neurological outcomes, with an unadjusted OR of 2.64 (95% CI, 1.39 to 5.00) and an adjusted OR of 2.37 (95% confidence interval [CI], 1.20 to 4.67).

Table 2.

Patient outcomes.

	Unadjusted					IPTW adjusted
	No bystander interventions (n = 175)	Bystander interventions				No bystander interventions (n = 196)	Abdominal thrusts (n = 189)		No bystander interventions (n = 248)	Back blows (n = 243)
		Abdominal thrusts (n = 24)		Back blows (n = 76)
	n (%)	n (%)	Differences [95% CI]	n (%)	Differences [95% CI]	n (%)	n (%)	Differences [95% CI]		n (%)	Differences [95% CI]
Primary outcome
Favorable neurological outcome	26 (15%)	9 (37%)	23% [3% to 43%]	24 (32%)	17% [5% to 29%]	30 (15%)	71 (38%)	22% [14% to 31%]	39 (16%)	75 (31%)	15% [8% to 23%]
Secondary outcome
Survival	62 (35%)	11 (46%)	10% [-11% to 32%]	45 (59%)	24% [11% to 37%]	68 (35%)	80 (42%)	8% [-2% to 17%]	88 (35%)	139 (57%)	22% [13% to 30%]
Prehospital cardiac arrest	139 (79%)	9 (38%)	−42% [-62% to –22%]	41 (54%)	−25% [-38% to −13%]	155 (79%)	116 (61%)	−18% [-27% to −9%]	194 (78%)	136 (56%)	–22% [-30% to −14%]

CI, confidence interval; IPTW, inverse probability of treatment weighting.

Patients who received back blows had a significantly higher survival rate and lower incidence of prehospital cardiac arrest compared to those who did not receive any bystander interventions (59% vs. 35%; difference 24%, 95% CI −11% to 32% for survival rate; 54% vs. 79%, difference −25%, 95% CI −38% to −13% for prehospital cardiac arrest). Cox proportional hazards model showed that back blows were independently associated with 30-day improved survival, with an unadjusted hazard ratio (HR) of 0.49 (95% CI, 0.33 to 0.73) and an adjusted HR of 0.52 (95% CI, 0.35 to 0.78). In contrast, abdominal thrusts were not significantly associated with improved survival, with an unadjusted HR of 0.75 (95% CI, 0.41 to 1.36) and adjusted HR of 0.73 (95% CI, 0.40 to 1.35) (Fig. 2).

Fig. 2.

Kaplan-Meier 1-month survival curves illustrating 1-month survival outcomes for patients who received bystander interventions compared to those who did not. (a) Comparison between abdominal thrusts and no intervention. (b) Comparison between back blows and no interventions. Patients who received back blows had a significantly higher survival rate compared to those who did not receive any bystander intervention. Twelve patients were excluded due to missing emergency medical services dispatch time.

Additionally, the success rate of abdominal thrusts and back blows were 30% and 27%, respectively. No significant differences were observed in the success rates between abdominal thrusts and back blows. One patient (4%) who received abdominal thrusts developed pneumothorax as a complication, whereas no complications were reported among patients who received back blows. The outcomes of the first and second attempts were presented in Supplementary Fig.B.

Discussion

This study represents the first large-scale research comparing the effectiveness of abdominal thrusts and back blows compared to no intervention. Our findings reveal that both abdominal thrusts and back blows were associated with favorable neurological outcomes compared to no intervention. Additionally, back blows were associated with higher survival rates.

Back blows

The ILCOR currently suggests that back blows should be used initially in patients with FBAO after ineffective coughing.⁸ However, the certainty of evidence was assessed as very low because the latest systematic review, which informs the current guidelines, found no intervention groups in the three studies on success rates and the one study on survival rates for back blows.10, 17, 18, 19 Consequently, it is not possible to quantitatively assess the effectiveness of back blows. Nevertheless, due to the intuitive, easy application, along with minimal complications, back blows are recommended as the initial intervention in several guidelines.8, 13, 20

This study demonstrated that back blows are associated with favorable neurological outcomes and higher survival rates, as well as significantly lower incidences of cardiac arrest, compared to no intervention. However, the success rate for the most severe cases was only 27%, with success rates in different settings varied from 10% to 72%.19, 21, 22, 23, 24 This discrepancy may be due to the fact that our registry only included patients who required transport to an emergency department by EMS. Cases in which foreign bodies were successfully removed at the scene without requiring hospital transfer were not captured, possibly leading to a selection bias toward more severe presentations. Therefore, back blows alone may be insufficient, and further research is needed on additional interventions and their sequencing.

Abdominal thrusts

First introduced by the American thoracic surgeon Henry Heimlich over half a century ago, the abdominal thrust has been globally recognized and advocated as a bystander intervention for FBAO.²⁵ Heimlich reported a 100% success rate with this maneuver.²⁶ Moreover, another study indicated that abdominal thrusts had a higher success rate than back blows (79% vs. 49%).¹⁹ As a result, there was a period in the 20th century when the American Heart Association (AHA) exclusively recommended abdominal thrusts as the initial intervention. However, these studies lack detailed data, and the evidence supporting the superiority of the abdominal thrusts over back blows remains insufficient. Furthermore, numerous fatal complications associated with abdominal thrusts have been reported such as aortic dissection, gastric rupture, and splenic rupture.27, 28, 29 Therefore, ILCOR currently suggests abdominal thrusts should be used if back blows are ineffective.⁸ The AHA continues to advocate for abdominal thrusts as the initial intervention due to its historical precedence.³⁰

The two studies on the success rate of abdominal thrusts and the six studies on survival rates all lacked control groups, making it impossible to comprehensively evaluate the efficacy of abdominal thrusts.18, 19, 31, 32, 33 However, this study revealed that abdominal thrusts are associated with favorable neurologic outcomes. Nevertheless, the success rate varies widely from 21% to 92%,19, 22, 23, 24 depending on patient severity, and the success rate for severe cases remains inadequate, similar to that of back blows.

Whereas both abdominal thrusts and back blows have been associated with favorable neurological outcomes compared to no interventions, further research is needed to determine the most effective first aid technique for initial intervention. The comparative effectiveness of abdominal thrusts and back blows remains unclear. A recent population-based cohort study revealed that back blows were significantly more effective for FBAO relief compared to abdominal thrusts, with an adjusted OR of 0.49 (0.30–0.80). While the majority of back blow cases originated from infants (less than one year old), subgroup analyses revealed that among individuals aged > 18 years and those in long-term care —populations most similar to ours in terms of age and frailty—back blows were more effective than other interventions in relieving airway obstruction. In Canada, where this study was conducted, back blows are recommended for individuals under one year old, and abdominal thrusts for those older than one year. Consequently, there were few cases where back blows were initially performed on adults. In Japan, where this study was conducted, the guidelines advocate back blows as the initial response, resulting in few cases where abdominal thrusts were performed first, leading to an insufficient number of cases.

Limitations

This study has several limitations. First, as an observational study, it is not possible to fully eliminate confounding, particularly due to unmeasured or missing variables. While we included key covariates in our propensity score model, several potentially important confounders—such as pre-event CPC, level of consciousness at the time of intervention, comorbidities, frailty, and witness characteristics—were either not collected or had substantial missing data. To address this, we used proxy variables such as baseline eating function, but residual confounding may remain. Second, this study evaluated only the first intervention performed by bystanders. We focused on the initial intervention because it is most relevant to real-time decision-making and guideline-based practice. Analyzing combinations would require complex time-dependent modeling and was limited by inconsistent reporting of timing and transitions between interventions. Further studies are needed to assess their cumulative effects. Third, cases where foreign bodies were successfully removed and did not require hospital transfer were not included in the registry, including those who refused transport after resolution by EMS or bystanders. Consequently, the results of this study may not be applicable to the overall FBAO population, including mild cases. Fourth, the cohort predominantly consisted of elderly individuals, many of whom were functionally dependent prior to the FBAO event. As a result, some patients may not have achieved favorable neurological outcomes due to pre-existing disability. To account for this, we adjusted for relevant confounders and reported survival as a complementary outcome measure. Fifth, given that the median age of all intervention groups exceeded 80 years and 75% of patients were older than 74 years, our findings primarily reflect outcomes in elderly populations and may not be generalizable to younger individuals.

Conclusion

This study demonstrates the crucial importance of bystander interventions for patients with FBAO. Both abdominal thrusts and back blows were associated with favorable neurological outcomes compared to no bystander intervention, with back blows additionally associated with higher survival rates.

Despite limitations such as the observational design and focus on severe cases, the findings support current guidelines and highlight the need for further research to optimize strategies for bystander intervention.

Conflicts of interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. This work was supported by the Japanese Association for Acute Medicine.

CRediT authorship contribution statement

Yutaka Igarashi: Writing – original draft, Visualization, Software, Resources, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Tatsuya Norii: Writing – review & editing, Validation, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Conceptualization. Ryuta Nakae: Writing – review & editing, Validation. Takashi Tagami: Writing – review & editing, Validation, Formal analysis. David P Sklar: Writing – review & editing, Validation, Conceptualization. Shoji Yokobori: Writing – review & editing, Validation, Supervision.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary material

The following are the Supplementary data to this article:

Supplementary Fig. A.

Distribution of multiple bystander interventions performed in cases of foreign body airway obstruction.

Supplementary Fig. B.

Propensity score distribution and covariate balance before and after inverse probability of treatment weighting. (a) Covariate balances between patients who received abdominal thrusts and who did not. (b) Covariate balances between patients who received back blows and who did not. The mean differences for each covariate are plotted for both unadjusted (red) and adjusted (blue) samples. The vertical dashed lines at -0.1 and 0.1 indicate thresholds for acceptable covariate balance. After adjustment, the covariates are more balanced, as most of the blue dots are closer to the zero line compared to the red dots, demonstrating the effectiveness of the adjustment methods used in the study.