A descriptive analysis of rural seniors with femur and pelvis fractures and the impact of morphine milliequivalents on inpatient rehabilitation

ABSTRACT

Background:

The US population ≥65 years old is rapidly growing. Geriatric disorders including hip fractures are becoming more numerous and the medical cost is significant. Hip fractures increase the risk of mortality in geriatric patients, but with improvements in medical care, this risk is declining over time. One important factor in the recovery and rehabilitation of hip fracture patients is pain management. Understanding the impact of opiate prescribing practices on outcomes in the rehabilitation setting can guide recommendations for the future.

Materials and Methods:

Patients ≥65 years old with hip fracture undergoing rehabilitation were retrospectively reviewed within the electronic health record. Information regarding gender, age, height, weight, BMI, length of stay, type of fracture, weight-bearing status, and comorbidities such as coronary artery disease, heart failure, chronic obstructive pulmonary disease, dementia, diabetes, renal disease, and vitamin D deficiency were obtained. The patients’ morphine milligram equivalents were calculated and averaged by length of stay. This information was analyzed with the GG score along with patient-specific information above to assess for a relationship using odds ratio and Wald confidence intervals.

Results:

Of the studied ≥65-year-old patients (n = 115), the amount of MME weakly positively correlated, with the change in GG score for MME/day (P = 0.01), mobility (P = 0.02) and self-care (P = 0.03) scores. Age >85 years old (P = 0.002) and BMI <25 (P = 0.046) correlated with lower MMEs/day. Negative correlation with GG improvement was seen for age (P = 0.002), dementia (P < 0.001), and avitaminosis D (P = 0.01). Age (P = 0.001), dementia (P < 0.001), and avitaminosis D (P = 0.03) also correlated negatively with mobility, while only dementia (P < 0.001) and avitaminosis D (P = 0.004) correlated negatively with an overall score. Weight-bearing status displayed the most consistent positive correlation with overall (P = 0.009) and mobility (P = 0.04) scoring. Isolated positive impacts were seen with ‘unknown’ fracture types on mobility (P = 0.04), femur fractures on self-care (P = 0.047), and pubic rami fractures on self-care (P = 0.04).

Conclusions:

A weak but significant correlation was seen between treatment with opioids and change in rehabilitation scoring. Age, dementia, and avitaminosis D correlated negatively in this rural, geriatric population, consistent with previous findings, while weight-bearing status and type of fracture had more positive correlations with functional scoring.

Introduction

In 2010, the US population ≥65 years old was projected to reach 20.9% by 2050.[1] Census results from 2020 reveal this population is already 16.0%, bringing with it an increase in geriatric syndromes.[2] While hip fracture incidence is impacted by the use of bisphosphates, overall numbers of hip fractures are projected to reach 840,000 by 2040 with an expected acute hospitalization cost of USD 6 billion.[3,4] Of these fractures, 86% occur in individuals ≥65 years old.[3,5] Studies show hip fractures are more costly than other osteoporotic fractures, nearly tripling a patient’s yearly cost of a fracture-free year.[6] When acute care and nursing home services are included, annual costs for hip fractures are USD 8 billion annually.[7] Neither estimate examines the loss of wages and other indirect costs of hip fracture endured by the family which can be substantial. Mortality costs of hip fractures can also be significant, however, 2005 Medicare claims show a decline in 1-year mortality rate from hip fracture to 21.9% [24.0% in 1986] for women and 32.5% [40.6% in 1986] for men.[3] This is thought to be related to changes in the management of hip fractures including timely surgical intervention, improved surgical devices, early weight-bearing exercise, interdisciplinary care, and increased discharge to rehabilitation centers rather than to home.[3,8]

Pain management is an important factor in the care of hip fractures and lack of control can lead to delirium, depression, delayed ambulation, pulmonary complications, and delayed transitions of care to lower levels of care including acute rehabilitation units.[9] It is well known that adequate pain control for various musculoskeletal disorders results in greater functional recovery, but studies use multiple techniques.[9,10,11,12] The techniques used to control pain are variable in the peri-operative period, but once patients are discharged to rehabilitation, physical therapy, and topical and oral analgesia are more commonly used. There has been much discussion in the literature on the prescription of opioids versus non-opioid analgesia.[12,13,14,15,16] Opiate prescribing practices are continuously changing, shaped in large part by the opioid epidemic in the United States.[17] Opioids have accounted for 80,816 overdose deaths in 2021, up from 70,029 in 2020.[18] While the geriatric population has been less effected by overdose deaths, they may be contributing to the epidemic by receiving and unintentionally distributing prescribed opioids.[14,19] Controlling pain in patients is a continuous struggle for physicians. Evidence suggests that the risks of opiates exceed the benefits of chronic pain treatment.[20] The effectiveness of opiates in acute pain is stronger, but its superiority to non-opioid pharmacotherapy is not conclusively proven.[12,16,21]

Rehabilitation goals in geriatrics include improved completion of activities of daily living and instrumental activities of daily living. Within the acute rehabilitation setting, this is measured by the GG score which was implemented in 2016 as a transition away from the functional independence measure (FIM) and other functional measures to a universally utilized tool in inpatient rehabilitation, skilled nursing facilities, home health, and long-term care hospitals.[22] Factors that may hinder rehabilitation goals have been previously assessed and include age, previous functional status, dementia, frailty, delirium, pre-operational nutritional status, depression, etc.[23,24,25] Uncontrolled pain can also contribute in many ways to reduced functional improvement following hip fracture.[9,11] Since pain control is often obtained with the use of opiates, the risks need to be weighed against the benefits. Risks of opiate use in geriatric patients can be greater than the younger population due to pharmacokinetic and pharmacodynamic changes associated with age.[26,27] Side effects include nausea, constipation, urinary retention, sedation, cognitive impairment, and dizziness which can affect over half of geriatric patients who use opiates.[27] To this point, no studies have been performed to describe a correlation between the amount of opiates measured in morphine milliequivalents (MME) and functional improvement during rehabilitation. This is important to guide clinicians in the use of opiates in rehabilitation of the older adults. Family medicine and other primary care providers often manage these patients in skilled settings, so knowing the factors, such as pain management, that correspond with greater success in functional recovery can improve outcomes. This study examines a retrospective correlation between MME and its effect on performance in rehabilitation, as measured by activities of daily living via GG, in older adults recovering from a hip fracture.

Materials and Methods

A retrospective review of the electronic health record was conducted to provide a descriptive analysis of rural seniors (≥65 years old) with femur and pelvis fractures and the impact of MME on inpatient rehabilitation. Demographic data obtained included gender, age, body mass index, and selected comorbidities: coronary artery disease, heart failure, chronic obstructive pulmonary disease, dementia, diabetes, renal disease, and vitamin D deficiency. Also collected were length of stay, type of fracture, and weight-bearing status. A patient’s daily MME was averaged, by summing the MME of the total stay and dividing it by the length of stay. Patients were evaluated with the GG per the practice of the institution and the Center for Medicare and Medicaid Services standardized functional assessment requirement. The GG is a standardized scoring system that evaluates activities of daily living (ADLs) and reports them as total, mobility scores and self-care scores.[22] Admission and discharge scores were obtained and analysis was conducted on the change between these GG scores. As this was a retrospective study, providers did not adjust their treatment practices.

Statistical analyses were performed using SAS9.4. A binary logistic regression model was built with the targeted variables of MME use. Predictor variables were age, BMI, gender, fracture side, fracture location, fracture type, weight-bearing status, and comorbidities of cardiac disease, lung disease, dementia, diabetes, renal disease, and vitamin D deficiency. The effects that were significantly associated with the target variable (P < 0.05) were identified and the odds ratio and their Wald confidence intervals were reported. Specific demographics were analyzed for their association with low (0–10 MME) or high (>10 MME) by simple t-tabled chi-square. The correlation was calculated by Pearson correlation. This project was approved by Marshall IRB (2027586-1).

Results

Demographic data for studied patients ≥65 years (n = 115) was categorized based upon average MME per day [Table 1]. Most patients fell into the either 0.1–10 MME/day (38 patients) and 10.1–20 MME/day (32 patients) groupings. Patients requiring no MME were the only group with fewer females than males (41.67%; P = 0.083) and a normal BMI (23.5; P = 0.046). Groups using less MME (0 MME/day and 0.1–10 MME/day) tended to be above 85 years and have a BMI lower than 25, as the amount of MME/day correlated negatively with age (P = 0.002) and positively with a normal BMI (P = 0.046). Lung disease was more prevalent in groups requiring higher MME/day (20.1–30 = 22.2% and >30 = 28.0%; P = 0.004).

Table 1.

Demographics of Studied Population Separated by MME

MME	Overall	0	0.1–10	10.1–20	20.1–30	>30	P
Number	115	12	38	32	16	17
Gender (percent female)	64.35%	41.67%	71.05%	65.63%	68.75%	28.82%	P=0.083
Age (years)	78.29	82.2	82.7	79.3	73.3	68.8	P=0.002
BMI	26.19	23.5	25.6	26	27.3	28.7	P=0.046
Type of fracture
Femur	101	75.00%	89.47%	90.63%	93.75%	82.35%
Pelvis	14	25.00%	10.53%	9.38%	6.25%	17.65%	P=0.528
Weight bearing status
Non-weight Bearing	7	8.33%	2.63%	9.38%	0.00%	11.76%
Paris weight bearing	13	8.33%	10.53%	12.50%	18.75%	5.88%
Weight bearing at time	95	83.33%	86.84%	78.13%	81.25%	82.35%	P=0.941
Comorbidities
CAD/CHF	30	16.67%	20.69%	20.00%	16.67%	20.00%
COPD	19	8.33%	8.62%	5.00%	22.22%	28.00%
Dementia	13	0.00%	12.07%	12.50%	0.00%	4.00%
Diabetes	28	16.67%	15.52%	25.00%	16.67%	16.00%
ERSD/Dialysis	10	0.00%	10.34%	5.00%	5.56%	4.00%
Vitamin D Deficiency	53	58.33%	32.76%	32.50%	38.89%	28.00%
Length of Stay (Days)	13.95	13.9	14.2	14.1	13	14
GG
Total
Admit	41.88	41.2	42	41.3	43.3	42.4
Change	54.58	53.5	48.9	53.7	59.8	64.8
Mobility
Admit	22.7	23.2	23.2	21.3	23.8	22.8
Change	38.41	27.6	37.6	38.5	41.9	45.9
Self-care
Admit	18.98	18.1	18.7	19.7	19.5	19.6
Change	15.56	15.9	14.1	15.1	17.9	18.9

While all groups scored similarly on all three GG entry scores, the amount of MME/day trended positively with the change in scores over the course of the treatment period. Pearson correlation measurements demonstrated low correlation for MME/day for change in overall (0.24; P = 0.01), mobility (0.23; P = 0.02), and self-care (0.20; P = 0.03) scores [Table 1].

Table 2 lists the impact of those demographics on the overall and specific components of the GG. Age correlates negatively with a decrease per year of age of 0.52 for the overall score (P = 0.002) and 0.44 for the mobility score (P = 0.001). The presence of dementia and vitamin D deficiency correlate negatively throughout all components of the score; with the former decreasing the overall score by 25.01 (P < 0.001), the mobility score by 18.51 (P < 0.001), and the self-care score by 7.59 (P < 0.001) and the latter decreasing the scores 9.19 (P = 0.009), 6.13 (P = 0.03) and 3.03 (P = 0.004) respectively. Relative positive impact on scoring was seen for weight bearing as tolerated in mobility scoring 10.40 (P = 0.04) and partial weight bearing across all three scores 36.53 (P = 0.004), 27.94 (P = 0.005), and 9.84 (P = 0.01). Weight-bearing status was the most consistent grouping, having a significant impact on overall (P = 0.009) and mobility (P = 0.04) scoring. Isolated positive impacts on score were seen with ‘unknown’ fracture types on mobility scoring (8.43; P = 0.04), femur fractures on self-care scoring (8.03; P = 0.047), and pubic rami fractures on self-care scoring (8.71; P = 0.04).

Table 2.

Correlation of MME with Rehabilitation Scoring

	Total GG	Mobility	Self-Care
MME/Day
Pearson	0.23562	0.22787	0.20486
P	0.0112	0.0148	0.0295
Total MME
Pearson	0.26518	0.20403	0.19932
P	0.0209	0.0294	0.0343

Discussion

Prescribing analgesic medications is a standard practice in the rehabilitation setting. The juxtaposition of the need for pain control and the risks both to an aging population and the presence of the opioid epidemic have questioned the practice of prescribing opioids versus non-opioid.[13,14,15] Very little in the literature has sought to discover correlations between the amount of opioid use MME and rehabilitation performance scores among older adults recovering from a hip fracture.

Within our population [Table 1], patients not requiring opioid analgesia during their rehab stay, tended to be less female (41.67%), as all other MME groupings were predominately female. That group was also the only one with a normal BMI (23.5), while the other groups were overweight which increased with each MME grouping. Age negatively trended with MME groupings similarly to a recent study reviewing the relationship between pain intensity and analgesic use in hip fracture patients.[28] Vitamin D deficiency was proportionally higher in the lower MME groupings, which is understandable as they were the oldest groupings. Finally, all three initial ratings of functional scoring were similar, demonstrating that the level of impairment was similar between groups. Changes in each aspect of the GG score (mobility, self-care, and overall) do positively trend with the amount of MMEs/day utilized [Table 1]. The literature does recognize associations with positive rehabilitation outcomes and patient ambulatory ability and ability to participate in physical and occupational rehabilitation.[8] Higher levels of pain preclude patient willingness or ability to participate in the tasks required to return to functional status. The higher levels of MME found in our results likely equates to better pain control, which allows better participation in the rehabilitation process. This correlation is not described in a study examining 163 patients recovering from unilateral trochanteric fractures of the femur as they found no association between analgesic use as (measured in MME) and functional outcomes (measured by GG scores),[28] but is described as a factor influencing mobility in a separate study on 143 post-surgical hip fracture patients.[29]

Other significant correlations are seen between GG scoring and the demographics of age, weight-bearing status, vitamin D, and dementia [Table 2]. As in Table 3, age correlates negatively with a decrease per year of age of 0.52 (P = 0.002) for the overall and 0.44 (P = 0.001) for the mobility scores. This correlates with prior studies indicating age as a poor prognostic factor regarding functional status after hip fracture.[24] The presence of dementia and vitamin D deficiency comorbidities correlate negatively throughout all components of the score. Decreases are seen by 25.01 (P < 0.001) in overall, 18.51 (P < 0.001) in mobility, and 7.59 (P < 0.001) in self-care scoring when dementia was present [Table 3]. Previous research has shown that individuals with dementia are 21 times more likely to be bedridden in the immediate post-operative period in comparison to their non-dementia counterparts which may be a factor in this negative association since the ability to bear weight seemed to correlate positively with improved functional status.[23] The diagnosis of vitamin D deficiency correlated with smaller, but still significant, decreases in the overall (9.19; P = 0.009), mobility (6.13; P = 0.03), and self-care (3.03; P = 0.004) scoring [Table 3]. Similarly, vitamin D supplementation has been examined in post-stroke rehabilitation scores measured in functional independent measures (FIM) and revealed improvements in comparison to placebo, but does remains controversial.[30] A positive correlation was found for partial weight-bearing status compared to non-weight bearing, having impacts of 36.53 (P = 0.009) points on overall, 27.94 (P = 0.04) on mobility, and 9.84 (P = 0.01) on self-care scoring [Table 3]. These agree with prior studies indicating patients with outside impediments to rehab participation, such as dementia or poor weight-bearing status, or indicators of frailty, such as increased age or vitamin D deficiency, perform worse and struggle to rehabilitate post-fracture.[8,23,24,30]

Table 3.

Impact of Demographic Factors on GG Scoring

Overall GG Change	Estimate	Wald 95% CI	Chi Square	P	Significant Finding Meaning
Age (years)	−0.52%	−0.85 to−0.20	10.09	0.002	Each year older decreases the score by 0.52
Boss mass index	−0.23	−0.75 to 0.29	0.75	0.39
Gender (female)	2.51	−4.76 to 9.78	0.46	0.50
Fracture site			1.07	0.59
Fracture Location			12.64	0.18
Fracture type			5.10	0.28
Weight bearing status vs. NWB
			16.03	0.009
PWB	36.53	11.54 to 61.52	8.21	0.004	PWB increases the score to 36.53
WBAT	11.29	−1.20 to 23.78	3.14	0.08
Comorbidities
Coronary/Heart failure	2.40	−5.29 to 10.08	0.37	0.54
Constructive lung disease	−2.66	−12.52 to 7.20	0.28	0.60
Dementia	−25.01	−35.39 to 14.64	22.33	<0.001	Dementia lowers the score to 25.01
Diabetes	−5.69	−13.36 to 1.99	2.11	0.15
ESRD/Dialysis	−2.79	−14.27 to 8.69	0.23	0.63
Vitamin D	−9.19	−16.03 to−2.36	6.95	0.009	Vitamin D deficiency lowers the score to 9.19
Mobility GG Change	Estimate	Wald 95% CI	Chi-Square	P	Significate Finding Meaning
Age (years)	−0.44	−0.70 to−0.19	11.73	0.001	Each year older decreases the score to 0.44
Body mass index	−0.23	−0.64 to 0.18	1.21	0.27
Gender (female)	2.74	−3.03 to 8.51	0.87	0.35
Fracture side			1.17	0.56
Fracture Location			12.75	0.17
Fracture type			6.14	0.19
Weight bearing status vs. NW
			17.47	0.04
PWB	27.94	8.33 to 47.56	7.79	0.005	PWB increases the score to 27.94
WBAT	10.40	0.59 to 20.21	4.32	0.04	WBAT increases the score to 10.40
Comorbidities
Coronary/Heart failure	3.28	−2.76 to 9.31	1.13	0.29
Constructive lung disease	−1.52	−9.58 to 6.55	0.14	0.71
Dementia	−18.51	−26.66 to−10.36	19.80	<0.001	Dementia lowers the score to 18.51
Diabetes	−5.49	−11.55 to 0.56	3.17	0.08
ESRD/Dialysis	−4.32	−13.33 to 4.69	0.88	0.35
Vitamin D	−6.13	−11.51 to−0.76	5.00	0.03	Vitamin D deficiency lowers the score 6.13
Self-care GG change	Estimate	Wald 95% CI	Chi-Square	P	Significant finding meaning
Age (years)	-0.08	−0.17 to 0.02	2.35	0.13
Body mass index	0.00	−0.15 to 0.16	0.00	0.97
Gender (female)	0.01	−2.20 to 2.18	0.00	0.99
Fracture site			1.36	0.51
Fracture Location			2.86	0.29
Fracture type			2.86	0.58
Weight bearing status vs. NWB
			10.30	0.07
PWB	9.84	2.30 to 17.39	6.54	0.01	PWB increases the score to 9.84
WBAT	1.00	−0.273 to 4.72	0.27	0.60
Comorbidities
Coronary/Heart failure	−0.53	−2.84 to 1.78	0.20	0.65
Constructive lung disease	−1.50	−4.57 to 1.57	0.92	0.34
Dementia	−7.59	−10.84 to−4.35	21.07	<0.001	Dementia lowers the score to 7.59
Diabetes	-0.07	−2.37 to 2.23	0.00	0.95
ESRD/Dialysis	0.87	−2.61 to 4.34	0.24	0.62
Vitamin D	−3.03	−5.07 to−0.98	8.41	0.004	Vitamin D deficiency lowers the score 3.03

Table 3 displays the impact that different factors have on the scoring for overall, mobility and self-care GG scoring. Significant findings within larger categories include an 8.43-point increase to mobility scoring of unknown fracture type. (P=0.04), and both an 8.03 point and 8.71-point increase in self-care scoring for femur fracture (P=0.047) and rami fracture

The main limitation of this data is that it is retrospective, which only allows the evaluation of what factors are associated with differing levels of opioid use during the rehabilitation stay but not causation. While understanding these associations is important, randomized studies would be necessary to determine if there is a causative relationship between higher MMEs and improved rehabilitation scoring. The correlation between MME/day and changes in all levels of GG scoring is significant, but low. This likely represents the relatively small sample size of this single-site study. It does provide a footstool to spark future research, which should focus on conducting this analysis in larger facilities of differing geographic region, size, and governance.

Conclusion

There has been recent momentum to decrease the prescribing practice of opioids. Our data, however, demonstrates there may be a place for some opioid analgesia in improvement in rehabilitation performance and return to functional status in patients with hip fracture. This likely represents the need for adequate analgesia for patients to perform optimally in rehabilitation efforts. Sometimes this analgesia may need to be in the form of opioids. To maintain responsible stewardship of opioid use, future research should target the relative efficacy of opioid analgesia versus other analgesic modalities.

Other correlations with functional outcomes seem to be related to age, vitamin D deficiency, dementia, and weight-bearing status. This has been previously demonstrated.[23,24,30] Rehabilitation potential is an important variable in the risk-benefit analysis of patient care. Identification of factors that can lead to hindrance and improvement of these goals could weigh heavily on decisions made by providers. The impact on pain control on these negative predictors could pose as an important topic for further studies.

Conflicts of interest

There are no conflicts of interest.

Funding Statement

Nil.