Association between post-extubation dysphagia and long-term mortality among critically ill older adults.

Abstract

Rationale.

Dysphagia following extubation is common in ICU patients. Diagnosing post-extubation dysphagia allows identification of patients who are at highest risk for aspiration and its associated adverse outcomes. Older adults are at an increased risk of post-extubation dysphagia and its complications due to multiple comorbidities, a higher baseline risk of dysphagia, and increased risk of pneumonia.

Objectives.

We aimed to investigate the association between post-extubation dysphagia and 1-year mortality in elderly patients. Secondary outcomes included ICU and hospital length of stay, ICU readmission, and place of discharge.

Methods.

We performed a retrospective cohort study from January 1 to December 31, 2013. ICU patients age 65 and over who were successfully extubated and underwent a formal swallow evaluation by a Speech and Language pathologist (SLP) were included. Dysphagia was graded using a 7-point scale and those with at least mild-moderate dysphagia were labeled as having clinically significant dysphagia.

Results.

Of 1075 patients who were screened, 359 were survivors age 65 and above, and of these 111 had a swallow evaluation performed by an SLP after liberation from mechanical ventilation. Mean age was 73.8 years (SD 7.0 years), and 41.4% had clinically significant dysphagia. In a multivariable regression model, there was no significant association between dysphagia and 1-year mortality. Furthermore, there was no statistically significant difference in ICU or hospital length of stay, ICU readmission, or place of discharge of those with clinically significant dysphagia compared to those without.

Conclusions.

Among mechanically ventilated ICU patients age 65 and over who underwent a swallow evaluation following extubation, dysphagia was not associated with mortality, ICU and hospital length of stay, ICU readmission, and place of discharge. Given conflicting evidence in the literature, larger prospective studies are needed to clarify whether post-extubation dysphagia is associated with worse outcomes in older patients admitted to the ICU.

Introduction

Patients who have undergone intubation and mechanical ventilation in an intensive care unit (ICU) are at risk for post-extubation dysphagia.^1–4 Endotracheal intubation itself can result in glottic injury and impair various swallowing protective reflexes.^3,5 Furthermore, various ICU therapies such as the use of sedatives and paralytics, and complications such as the development of gastroesophageal reflux can predispose to alterations in swallowing and lead to aspiration.^5,6 Finally, ICU-acquired weakness can affect not only strength in proximal extremities and the muscles of respiration, but this syndrome likely affects oropharyngeal muscle strength as well.⁷ Although post-extubation dysphagia is a well-recognized phenomenon, particularly in those who have been ventilated longer than 48 hours, estimates of its incidence vary widely in literature, from as low as 3% to as high as 83%.⁸ This variability is in part affected by differences in the evaluation for dysphagia, which can include clinical assessment at the bedside by a nurse or speech and language pathologist (SLP), or instrumental evaluation using fiberoptic endoscopic evaluation of swallow (FEES) or radiographic assessment via video fluoroscopic swallowing examination. Advanced age, the presence of chronic conditions prior to critical illness such as congestive heart failure (CHF), stroke, and neuromuscular weakness have been shown to be risk factors for post-extubation dysphagia, as have factors during critical illness including the duration of mechanical ventilation.^2,9–11 In addition, advanced age is a risk factor for slower resolution of post-extubation dysphagia.¹²

The association between older age and a higher risk of dysphagia and aspiration is multifactorial. Increasing age is associated with physiologic changes in swallow function, and age-related conditions such as stroke and cognitive impairment are significant risk factors for the presence and severity of dysphagia and aspiration.^13–15 Older adults are also at increased risk of developing pneumonia caused by aspiration. Contributors to this risk include increased bacterial burden in the oropharynx, particularly those who have been hospitalized or institutionalized,^16,17 as well as age-related changes in cellular and humoral immunity.^18,19 With the aging of the population and an ongoing trend towards increasing utilization of intensive care for older adults, post-extubation dysphagia may become more frequently encountered.^20–22 Therefore there is interest in post-extubation dysphagia as a risk factor for pneumonia and recurrent respiratory failure, which might convey risk of worsened short-term as well as long-term outcomes.²³

There are other reasons why post-extubation dysphagia may be associated with long-term outcomes. Increased interest in and knowledge of long-term outcomes after critical illness such as cognitive and physical functional outcomes ^24–28 has led to increased investigation into the ability to predict long-term outcomes during critical illness.^29–31 While there are no clinical assessments of global cognition or strength that are routinely performed in critically ill patients, rigorous evaluation for dysphagia following extubation is quite common. Deconditioning and frailty may be associated with dysphagia.³² The fact that weakness of the tongue, one important contributor to dysphagia in older adults, is associated with weakness of grip strength supports the hypothesis that dysphagia might be a useful marker of frailty.^33–35 Frailty is a common and important clinical syndrome in older adults characterized by age-associated declines in physiologic reserve and function across multiple organ systems, leading to increased vulnerability for adverse health outcomes including death. Thus it is possible that post-extubation dysphagia is associated with poor long-term outcomes following critical illness, beyond the short-term risks of adverse outcomes such as recurrent respiratory failure and aspiration pneumonia. One more recent longitudinal study of ARDS survivors showed that one third of those who were intubated suffer dysphagia after discharge, with length of time to recovery tied to their ICU length of stay.³⁶

We hypothesized that post-extubation dysphagia in critically ill adults aged 65 years and older would be associated with higher one-year mortality, both because of risks directly imposed by the presence of dysphagia and because dysphagia might serve as a marker of frailty and generalized muscle weakness. We sought to investigate this hypothesis by performing a retrospective cohort study of elderly patients who were successfully extubated in an ICU in an academic tertiary hospital and who survived to hospital discharge. We also investigated outcomes of hospital length of stay, ICU readmission, and discharge destination.

Methods

Study Design

We completed a retrospective cohort study of patients aged 65 years and older admitted to a 24-bed academic medical and surgical ICU between January 1, 2013 and December 31, 2013 who had been intubated, were successfully extubated, and survived to hospital discharge. Clinical data were obtained from electronic health records of the University of Wisconsin Hospital and Clinics. The study was approved by the University of Wisconsin Institutional Review Board.

Inclusion/Exclusion Criteria

Patients aged 65 and over who had been successfully liberated from mechanical ventilation via endotracheal tube, who subsequently underwent evaluation by a Speech-Language Pathologist following extubation as part of their clinical care and survived to hospital discharge were included in this cohort. Exclusion criteria included the placement of tracheostomy either before admission or during the hospital stay, admission with head and neck injuries, traumatic brain injury, or for a primary CNS condition including stroke, intracranial hemorrhage, subarachnoid hemorrhage, and neuromuscular disease. Patients were also excluded if they were extubated as part of withdrawal of life-sustaining therapy. Evaluation for post-extubation dysphagia included some combination of bedside swallow evaluation, fiberoptic endoscopic evaluation of swallow (FEES), and videofluoroscopic swallow (VFS) as determined by the Speech-Language pathologist.

Data Collection

Patients who were mechanically ventilated were identified using a hospital quality assurance database. Baseline demographic data, comorbidities, were obtained from the University of Wisconsin Hospital and Clinics Electronic Health Record (EHR) software (Epic Systems, Verona, WI). APACHE IV score was prospectively calculated upon ICU admission as part of routine quality assurance.³⁷ The results of evaluations performed by the Speech-Language pathologist were obtained from consult and progress notes in the EHR. Dysphagia was graded using a 7-point scale ranging from none to severe dysphagia, in parallel with the American Speech-Language-Hearing Association Functional Communication Measure for swallowing (FCM 1–7). ³⁸ The presence of penetration or aspiration was recorded for those patients who were clinically assessed with VFS or FEES.

The primary outcome was 1-year mortality. We used the EHR as well the social security death index (SSDI) to ascertain vital status 1 year after hospital discharge. ³⁹ Secondary outcomes were hospital length of stay, discharge destination, and reintubation. The rate of reintubation was obtained from instances where an endotracheal tube was reinserted within 30 days of extubation after a swallow evaluation was performed.

Statistical Analysis

Dysphagia was dichotomized, with those with an ASHA FCM level 4 and below for swallowing (at least mild-moderate dysphagia) was grouped in one category and compared to those with FCM level 5–7 (at most mild dysphagia). Those with at least mild-moderate dysphagia were classified under “clinically significant dysphagia” for purposes of the analyses. Unadjusted binary outcomes (presence of significant dysphagia, mortality) were assessed with χ2 test. The associations between discharge to a skilled nursing facility, reintubation, and the presence of significant dysphagia were also analyzed with χ2 test. The association between hospital length of stay and clinically significant dysphagia was assessed with a two-sample Wilcoxon rank-sum (Mann-Whitney) test. Finally, the association between dysphagia and 1-year mortality was assessed with multivariable logistic regression, adjusting for prespecified potential confounders of age at admission, sex, severity of illness on the first ICU day, dummy variables for each category of chronic conditions, body mass index at ICU admission, and medical versus surgical patients.

Results

During the period from January 1, 2013 to December 31, 2013 a total of 1075 patients were admitted to the ICU and mechanically ventilated. Records of the 468 subjects who were age 65 and over at the time of ICU admission were reviewed. Those who did not survive to hospital discharge (109, or 23% of this group)), with an admission diagnosis of a stroke or intracranial hemorrhage (16, 3.4%), trauma to the head or neck (21 4.5%), or with tracheostomy either present during admission or performed during the hospitalization (13, 2.8%) were excluded from the study. Of the remaining 309 patients who survived to hospital discharge, 111 (35.9%) had undergone evaluation by a Speech-Language pathologist in the ICU following extubation, based upon a decision of the attending physician to request Speech-Language Pathologist consultation. These 111 patients were included in the analyses. Figure 1 summarizes the study design. Mean age of the cohort was 73.8 years (standard deviation, SD, 7.0 years), 41.4% were women, the median duration of mechanical ventilation was 2 days (Interquartile range, IQR, 1 to 4 days), and the median length of stay in the ICU was 4 days (IQR 2–6 days). The mean APACHE IV score was 87.1 (SD 25.6), corresponding to in-hospital mortality probability of approximately 62%. Thirty-six percent of the subjects were intubated in the operating room, while 26% were intubated in the ICU, with the remaining 35% intubated in the field, in the emergent department, or in another hospital prior to transfer. Prehospital comorbid conditions are also listed in Table 1.

Figure 1.

Flowchart of the formation of the retrospective cohort

TABLE 1.

DEMOGRAPHIC CHARACTERISTICS OF STUDY POPULATION

	Significant Dysphagia
	absent (n=65)	present (n=46)	p
DEMOGRAPHICS
AGE	73.8	73.9	0.97
FEMALE SEX	28 (43.1%)	18(39.1%)	0.88
RACE (CAUCASIAN)	62 (95.45%)	46 (100%)	0.14
BMI, Μ (Σ)	31.6 (9.5)	28.7 (6.9)	0.08
COMORBIDITIES
	13 (20.0%)	5 (10.9%)	0.20
CEREBROVASCULAR DISEASE
COPD	18 (27.7%)	12 (26.1%)	0.85
OSA	6 (9.2%)	3 (6.5%)	0.61
CHF	25 (38.5%)	18 (39.1%)	0.94
CHRONIC LIVER DISEASE	3 (4.6%)	6 (13.0%)	0.10
RENAL FAILURE	15 (23.1%)	8 (17.4%)	0.47
MALIGNANCY	15 (23.1%)	13 (28.3%)	0.54
APACHE IV, Μ (Σ)	86.6 (30.0)	87.7 (25.4)	0.82
LOCATION OF INTUBATION			0.24
PRE-HOSPITAL	0	3 (6.5)
ED	6 (9%)	7 (15%)
ICU	19 (29%)	10 (21%)
OPERATING ROOM	26 (40)	14 (30)
OTHER	14 (22%)	12 26%)
DURATION OF INTUBATION (D, MEDIAN)	2 (IQR 1–3)	2 (IQR 1–4)	0.10
MULTIPLE INTUBATIONS	16 (24.6%)	9 (19.6%)	0.53
CLINICAL SERVICE
MEDICAL	38(58%)	32(70%)	0.23
SURGICAL	27(42%)	14(30%)
INSTRUMENTAL SWALLOW EVALUATION PERFORMED	28 (43%)	35 (76%)	0.001

The Speech-Language Pathologist evaluating the patient decided whether additional evaluation with either FEES or videofluoroscopy were indicated based upon clinical evidence of aspiration upon bedside examination. Forty-seven (42%) patients underwent clinical evaluation only, 29 (26%) patients underwent FEES, 32 (29%) were evaluated with videofluoroscopy, and 2 patients underwent evaluation with both FEES and videofluoroscopy. Dysphagia severity was determined by the Speech-Language Pathologist based on the findings of the clinical or clinical and instrumental evaluation that was performed. Forty-six out of 111 (41.4%) patients had clinically significant dysphagia (at least mild-moderate dysphagia), while the remaining 65/111 (58.6%) did not have clinically significant dysphagia.

Dietary modifications resulting from evaluation by the Speech-Language pathologist were also recorded. In those with significant dysphagia, about 63% were treated with diet modification alone and 37% were placed on nil per os (NPO) status. Dietary modifications included administration of dysphagia diet or changes in the consistency of liquids. In contrast, in those without significant dysphagia, only 3% were placed on NPO, and 40% were treated with some form of dietary modification.

There was no statistically significant difference in the primary outcome, 1-year mortality, between those with clinically significant dysphagia compared to those without (Figure 2). One-year mortality was 35% in those with clinically significant dysphagia, and 32% in those without (p=0.638). There was no statistically significant difference in ICU readmission between those with and without clinically significant dysphagia. Discharge to a long-term acute care hospital or a skilled nursing facility occurred in 42% of those with clinically significant dysphagia compared to 44% in those without clinically significant dysphagia, again with no statistically significant difference (p=0.858). Moreover, there was no significant difference in median ICU or hospital length of stay in those with clinically significant dysphagia compared to those without (Figure 2). Finally, in a multivariable regression model adjusting for age at admission, sex, severity of illness on the first ICU day, comorbid conditions, body mass index at ICU admission, and medical versus surgical patients, there was no significant association between the presence of at least mild-moderate dysphagia and 1-year mortality (odds ratio 0.73, 95% confidence interval 0.28 to 1.92, p=0.53) (Table 2).

Figure 2.

Key clinical outcomes of ICU and hospital length of stay (shown with box plots in A), the proportion readmitted to the ICU (B), the proportion discharged to home (C), and 1-year mortality (D) did not differ by the presence of significant dysphagia on evaluation. The association between dysphagia and length of stay was evaluated with the two sample Wilcoxon rank-sum (Mann-Whitney) test, and between dysphagia and binary variables with the χ2 test.

Table 2.

Association^a between at least moderate dysphagia and the odds of death at 1 year, multivariable logistic regression

	Odds Ratio	95% Confidence Interval	p
Dysphagia	0.73	0.28 to 1.92	0.53
Age	1.01	0.94 to 1.1	0.76
Male sex	0.60	0.24 to 1.53	0.29
APACHE IV score	1.01	0.99 to 1.03	0.19
Medical patient	2.73	0.99 to 7.6	0.05
Hospital length of stay	1.04	1.00 to 1.08	0.02
BMI	0.94	0.88 to 0.99	0.02
History of comorbid illness
Cerebrovascular disease	0.65	0.18 to 2.39	0.52
Congestive heart failure	1.20	0.42 to 3.42	0.74
COPD	0.75	0.25 to 2.24	0.60
Obstructive sleep apnea	0.19	0.02 to 1.62	0.13
Malignancy	2.04	0.75 to 5.57	0.16
Liver disease	1.09	0.11 to 10.4	0.94
Chronic kidney disease	2.02	0.58 to 7.12	0.27
Diabetes mellitus	1.48	0.46 to 4.77	0.51
Chronic neurologic disease	0.34	0.88 to 0.99	0.02

^aAssessed with multivariable logistic regression, adjusting for prespecified potential confounders of age at admission, sex, severity of illness on the first ICU day, dummy variables for each category of chronic conditions, body mass index at ICU admission, and medical versus surgical patients.

Discussion

Clinically significant dysphagia following extubation was common in this cohort of older medical and surgical ICU patients liberated from mechanical ventilation who survived to hospital discharge, being seen in over 40% of these patients. Our study did not demonstrate an association between dysphagia of at least moderate severity and adverse outcomes including 1 year-mortality, ICU readmission, or longer hospital stay. These findings do not support the idea that post-extubation dysphagia is a useful surrogate for other functional measures in a way that could be useful for predicting long-term functional outcomes in older survivors of critical illness.

These results contrast with those of a larger retrospective study performed in medical and surgical ICU patients, which demonstrated that post-extubation dysphagia was associated with higher rates of reintubation, pneumonia, and in-hospital mortality.²³ There are important differences between these studies. We were interested in the association between dysphagia and long-term mortality among hospital survivors. Thus while post-extubation dysphagia may be significantly and even causally related to higher short-term mortality, the results of these studies taken together suggest that among those patients who survive this higher risk period there is no important relationship between post-extubation dysphagia and longer term outcomes. Moreover, a significant number of patients in our study underwent instrumental evaluations with videofluoroscopy and FEES. Instrumental evaluations of swallow function are more objective and likely more sensitive in assessing for dysphagia.⁴⁰ On the other hand, the determination of penetration or aspiration in a single study might not necessarily translate to the development of adverse outcomes later during or following hospitalization.

A prospective study of patients in a mixed adult ICU also showed an association between post-extubation dysphagia and 90-day mortality.⁴¹ In contrast to our study, a significant proportion of patients who had dysphagia after ICU discharge were admitted for neurologic disease. Our study had excluded this subset of patients with high incidence of dysphagia independent of their intubation status. Interestingly, the study also followed the evolution of dysphagia and recovery, and found that recovery from dysphagia was lowest in patients admitted for neurologic indications as compared to patients with other reasons for admission.

In a previous study of elderly patients in a medical intensive care unit in which aspiration occurred in just over half of those age 65, 40% had persistent swallowing dysfunction after 5 days but only 14% after 2 weeks.¹² These data suggest that post-extubation dysphagia resolves relatively quickly in the vast majority of patients. There is further evidence that alteration in the swallowing reflex after prolonged intubation improves after 1 week,⁴² and that even in cohorts where the prevalence of post-extubation aspiration on FEES is over 50%, clinically significant aspiration events while in the hospital may be uncommon.⁴ The fact that our study identified no associations among those who survive to hospital discharge suggests that while post-extubation dysphagia may convey risks in the short term, patients who emerge from this period without complications or consequences of dysphagia are not at significantly higher risk of poor outcomes.

It might be tempting to conclude that our results call into question the value of evaluation for post-extubation dysphagia. We do not think our findings support this conclusion. The detection of clinically significant dysphagia led to further clinical evaluation and for most subjects some form of diet modification. In those with significant dysphagia, over half were treated with diet modification alone and just over 1/3 were placed on nil per os (NPO) status. Our study cannot determine the efficacy of such interventions following evaluation for dysphagia, but it is reasonable to think that these may mitigate the short-term risks of aspiration and associated complications in the early post-extubation time period. However, there is not a robust evidence base to guide the choice of evaluation for dysphagia following extubation and thus the benefits of any one approach relative to others remain unknown.

The study has a number of limitations. First, we identified our cohort retrospectively from a convenience sample of elderly mechanically ventilated patients who were referred for a swallow evaluation. The population studied may represent a subgroup who are at a high risk for acquiring post-extubation dysphagia or aspiration, and we would assume that the prevalence of significant dysphagia would be higher in this population than when considering all older patients who have been extubated. However, given the non-systematic way that patients were referred for evaluation by a Speech-Language Pathologist (i.e. the clinical judgement of the critical care physician) we cannot investigate this question in our cohort and this may limit the generalizability of our findings. Second, in our multivariable regression model investigating the association between dysphagia of at least moderate severity and 1-year mortality, our confidence intervals are quite wide and allow for the possibility that a significant positive (or even negative) association truly exists. This speaks to the limitations in our sample size. Thirdly, our study cannot differentiate premorbid dysphagia related to age-related changes or chronic conditions from dysphagia which developed as a consequence of endotracheal intubation from laryngeal injuries including edema, ulceration, and granulation tissue formation which can result in dysphagia after extubation.⁴³ Given the lack of significant associations detected in our study, this may be less of an issue. Finally, there likely was variability in the swallow assessments of the Speech and Language pathologists who performed the evaluations. The interpretation of swallow function can also be influenced by the method of evaluation, whether by clinical or instrumental evaluation. This could have impacted the measurement of dysphagia severity and the choice of instrumental tests of swallow. Furthermore, the interpretation of swallow function can also be influenced by the method of evaluation, whether by clinical or instrumental evaluation.

In conclusion, we did not find any significant association between dysphagia and 1-year mortality among elderly patients age 65 and over who have been successfully liberated from mechanical ventilation and subsequently underwent a swallow evaluation by a Speech and Language pathologist. In addition, the presence of clinically significant dysphagia did not impact outcomes such as ICU readmission, ICU and hospital length of stay, and place of discharge. The results of our study suggest that post-extubation dysphagia may not be predictive of long-term outcomes such as mortality in elderly patients admitted to the ICU. Further investigation with a larger sample size is warranted to elucidate the role and effects of post-extubation dysphagia in this patient population, and evaluate the effectiveness of various therapies in mitigating the adverse outcomes of dysphagia and aspiration.