Abstract
Background
chronic pain, a common complaint among older adults, affects physical and mental well-being. While opioid use for pain management has increased over the years, pain management in older adults remains challenging, due to potential severe adverse effects of opioids in this population.
Objective
we examined the association between opioid use, and changes in cognitive function of older adults.
Design
prospective study.
Setting
community dwelling older adults.
Subjects
study population consisted of 2,222 individuals aged 65–69 years at baseline from the Personality and Total Health Through Life Study in Australia.
Methods
medication data were obtained from the Pharmaceutical Benefits Scheme. Cognitive measures were obtained from neuropsychological battery assessment. Opioid exposure was quantified as Total Morphine Equivalent Dose (MED). The association between change in cognitive function between Wave 2 and Wave 3, and cumulative opioid use was assessed through generalized linear models.
Results
cumulative opioid exposure exceeding total MED of 2,940 was significantly associated with poorer performance in the Mini Mental State Examination (MMSE). Compared with those not on opioids, individuals exposed to opioids resulting in cumulative total MED of greater than 2,940 had significantly lower scores in the MMSE (Model 1: β = −0.34, Model 2: β = −0.35 and Model 3: β = −0.39, P < 0.01). Performance in other cognitive assessments was not associated with opioid use.
Conclusion
prolonged opioid use in older adults can affect cognitive function, further encouraging the need for alternative pain management strategies in this population. Pain management options should not adversely affect healthy ageing trajectories and cognitive health.
Keywords: cognition, cognitive ageing, opioid, opioids, pain, older people
Key points
Long-term opioid use significantly affects the cognitive function of older adults.
Presence of Apolipoprotein E (APOE) epsilon 4 (ε4) genotype increases the risk of adverse cognitive effects of opioid use.
Pain management using opioids must be approached with caution in older adults.
Introduction
Older adults commonly complain about persistent pain. Among older adults in nursing homes, persistent pain is reported in ~80% of the population. Close to 50% of community dwelling older adults report similar complaints [1]. Chronic pain in older adults is linked to impaired physical function and psychological distress [2]. Pain management in this population has its unique challenges due to age-related risks of adverse effects as well as the high prevalence of polypharmacy among older adults [3].
While it is not recommended as the first choice of chronic pain management in older adults, opioids are being prescribed more often now and for longer period of use among older adults [4–7]. At least one prescription of opioid is filled in a year by 18% of adults in the USA [8]. In Australia, 22% of community dwelling older adults report regular opioid use [9]. Among older adults, multiple adverse effects of opioid use have been reported. These include side effects such as nausea, constipation, urinary retention, sedation and dizziness [10–12]. When used concomitantly with antipsychotics, opioids increase the risk of falls in older adults [13]. Opioids are also associated with delirium [14] and have been reported to result in cognitive impairment as well as hallucinations in older adults [15, 16].
Chronic exposure to opioids can introduce tolerance and more seriously, dependence to the drugs [17]. Tolerance is a result of a desensitization mechanism and will result in the need for higher doses to produce the desired effect. This is particularly dangerous in the older adults, due to dose–response related adverse events [17]. Opioids are also shown to affect cellular biology in the brain, which can promote the development and progression of neurodegenerative diseases [18].
This study examined the association between prescription opioid use in older adults and changes in cognitive function over a period of 4 years. We also examined the effect modification of this association by the apolipoprotein E epsilon 4 (APOE-ε4) gene.
Methods
Study design
The Personality and Total Health (PATH) Through Life Study aimed to assess the course of diseases over the lifespan of adults living in the community through a longitudinal cohort study design. Conducted among Australian adults, the PATH study sample consisted of three separate cohorts. To date, the PATH study has followed the three cohorts of participants starting at ages 20–24, 40–44 and 60–64 years for ~20 years. Age cohorts were narrow with birth years 1975–79, 1956–60 and 1937–41 for the 20+ cohort, the 40+ cohort and the 60+ cohort, respectively. Participants were randomly selected from the electoral rolls of the Australian Capital Territory and Queanbeyan, Australia. Inclusion criteria include participants with a residential address in Queanbeyan and within the age range for the specific cohorts at baseline (20–24, 40–44 and 60–64 years).
Starting between the year 2000 and 2002, the study cohorts were followed at 4-year intervals. Starting with the 20+ cohort, each cohort was interviewed successively. Participation rates for follow-up visits across the cohorts ranged from 89 to 93%. Three sub-studies derived from sub-samples of the main PATH study have been done. These studies are the Magnetic Resonance Imaging study, the Health and Memory study and the Cardiovascular study. The PATH study design and study population were previously described [19]. The PATH study was approved by the Australian National University Ethics Committee. Participants provided informed consent to participate in the study and to have their data linked to the Australian Government Pharmaceutical Benefit Scheme (PBS) data.
Study population
This study focuses on the 60+ cohort of the PATH Through Life study, from Wave2 to Wave 3 of the PATH study. Baseline data for this study were obtained from the first follow-up wave of the PATH study (Wave 2, year 2005/2006, n = 2,222) and 4-year follow-up data were obtained from the second follow-up wave of the PATH study (Wave 3, year 2009/2010, n = 1973). Baseline covariates and baseline cognitive function was obtained from Wave 2. Follow-up cognitive function was obtained from Wave 3.
Exposure measure
Medication data for this study were obtained from the Pharmaceutical Benefit Scheme (PBS). The PBS and pharmaco-epidemiology research done using PBS is described elsewhere [20, 21]. The cumulative total morphine equivalent dose (MED) of opioids used between Wave 2 and Wave 3 of the PATH Study was used to quantify the exposure measure in this study. Medications listed as opioids according to the World Health Organization Anatomical Therapeutic Chemical classification system were included in this study [22]. The total MED for each opioid prescription was derived by first computing the product of the quantity of opioids by their respective strengths. This product was then multiplied with medication-specific morphine conversion factors [23, 24]. For each study participant, total MED for all opioids taken between Wave 2 and Wave 3 of the PATH study was summed to derive the cumulative total MED. The list of opioids used by participants and their respective morphine conversion factors in this study is presented in Table 1.
Table 1 .
|
Natural opium alkaloids • Codeine phosphate hemihydrate (0.15) • Morphine sulfate pentahydrate (1.00) • Morphine hydrochloride trihydrate (1.00) • Oxycodone hydrochloride (1.50) • Hydromorphone hydrochloride (4.00) • Oxycodone hydrochloride + naloxone hydrochloride (1.50) Diphenylpropylamine derivatives • Methadone hydrochloride (3.00) Phenylpiperidine derivatives • Fentanyl (7.20) Oripavine derivatives • Buprenorphine (12.60) Opioids in combination with non-opioid analgesics • Paracetamol + codeine phosphate hemihydrate (0.15) Other opioids • Tramadol hydrochloride (0.10) • Tapentadol (0.40) |
Outcome measures
Outcome of interest in this study is change in cognitive function from Wave 2 to the next 4-year follow-up (Wave 3). Multiple cognitive domains were assessed through a series of neuropsychological tests. The Mini Mental State Examination (MMSE) was used to assess global cognition [25]. Short-term memory was assessed through Immediate Recall and Delayed Recall using the California Verbal Learning Test [26]. The Wechsler Memory Scale-Digit Span Backward was used to assess working memory [27] and the Symbol-Digit Modalities Test [28] was used to assess information processing. Verbal ability was assessed with Spot-the-Word Task [29], while psychomotor speed and information processing were assessed with Simple Reaction Time and Choice Reaction Time [30]. Trail Making Test, parts A and B were used to assess processing speed [31] and executive function, while Purdue Pegboard Test was used to assess psychomotor speed [32].
Covariates
Multiple covariates linked to opioids and cognitive function were included in the analysis. These included demographic factors such as age, sex and years of education. Self-reported smoking status (coded as current smoker, past smoker and never smoked), alcohol consumption (assessed as drinks per week) and physical activity (categorized as hours of mild, moderate and vigorous activity) were also included in the analysis. Clinical risk factors such as self-reported stroke, diabetes and family history of dementia, as well as depression assessed with the Patient Health Questionnaire (PHQ-9), were included as covariates in the final model. Additionally, hypertension status (defined as systolic blood pressure ≥ 140 mmHg, diastolic blood pressure ≥ 90 mmHg or self-reported use of antihypertensives) and body mass index (BMI) (computed as weight (kg)/height (m)2) were also included. Apolipoprotein E epsilon 4 (APOE-ε4) genotype was added as a covariate of interest in this analysis and used to assess effect modification. Genotyping for APOE variants in the PATH study population is described elsewhere [33].
Statistical analysis
Data analysis was completed in SAS (v.9.4, SAS Institute Inc., Cary, NC). Demographic and health related characteristics for the PATH study population at Wave 2 were examined with bivariate analysis using t-tests and Fisher’s exact test. Generalized Linear Models (GLMs) were used to assess the association between opioids and change in cognitive function, and the effect modification of this association by APOE-ε4 allele. Three statistical models, representing an unadjusted model (Model 1), a partially adjusted model for demographic variables (Model 2) and a fully adjusted model with all covariates (Model 3) were used to assess these associations. To minimize Type 1 error due to multiple comparisons, statistical significance was maintained at 0.01.
Results
Population characteristics
Table 2 describes the baseline study population characteristics based on opioid exposure. Overall, of the 2,222 individual from Wave 2, 1973 (88.8%) individuals completed all study assessments at Wave 3. Of the 249 (11.2%) of individuals who were not assessed at Wave 3, 136 (6.1%) individuals refused participation at Wave 3 and 11 (0.5%) were not found and 102 (4.6%) died between Waves 2 and 3. In this study, 440 individuals (19.8%) were prescribed opioids during the study period. Individuals on opioids had lower total years of education (13.3 compared with 14.1 years) and higher BMI (28.3 compared with 26.4). Compared with those not on opioids, a higher percentage of individuals who were on opioids were not in the labor force (79.3 compared with 72.3%) and were past smokers (41.5 compared with 36.3%) or currently smoking (15.3 compared with 8.3%). The percentage of participants with a history diabetes was higher in the group exposed to opioids. The percentage of individuals with depression was also higher among those exposed to opioids compared with those not exposed to opioids (16.3 compared with 8.4%).
Table 2 .
Baseline characteristics of study participants according to opioid use: PATH Through Life Study, Wave 2, 2003–2006
| Variable | Not using opioids (N = 1,782) |
Using opioids (N = 440) |
|---|---|---|
| Demographic factors Age (years), mean (std) |
66.6 (1.5) | 66.6 (1.5) |
| Gender, n (%) Male Female |
935 (52.5) 847 (47.5) |
212 (48.2) 228 (51.8) |
| Race, n (%) White Aboriginal/Torres Strait Islander Asian Other |
1,703 (95.6) 1 (0.1) 47 (2.6) 30 (1.7) |
430 (97.9) 0 (0) 5 (1.1) 4 (0.9) |
| Education (years), mean (std)*** | 14.1 (2.7) | 13.3 (2.6) |
| Marital Status, n (%) Married Unmarried-living with partner Separated Divorced Widowed Never married |
1,328 (74.6) 64 (3.6) 36 (2.0) 168 (9.4) 142 (8.0) 43 (2.4) |
314 (71.4) 12 (2.7) 9 (2.0) 52 (11.8) 46 (10.5) 7 (1.6) |
| Employment status, n (%)** Employed full-time Employed part-time, looking for full-time employment Employed part-time Unemployed, looking for work Not in the labor force |
180 (10.1) 1 (0.1) 310 (17.4) 2 (0.1) 1,288 (72.3) |
17 (3.9) 1 (0.2) 72 (16.4) 1 (0.2) 349 (79.3) |
| Lifestyle factors | ||
| Smoking status, n (%)*** Never smoked Past smoker Current smoker |
986 (55.4) 647 (36.3) 148 (8.3) |
190 (43.3) 182 (41.5) 67 (15.3) |
| Alcohol consumption (drinks per week), mean (std) | 6.8 (8.4) | 7.0 (8.8) |
| Physical activity (hours per week), mean (std) Mild activity Moderate activity Vigorous activity |
7.8 (8.9) 2.8 (4.7) 0.8 (2.5) |
8.2 (9.4) 2.6 (4.8) 0.6 (2.1) |
| Clinical factors History of stroke, n (%) Yes No |
44 (2.5) 1,698 (97.5) |
18 (4.2) 412 (95.8) |
| History of diabetes, n (%)** Yes No |
155 (9.0) 1,564 (91.0) |
63 (14.9) 361 (85.1) |
| History of hypertension, n (%)* Yes No |
607 (34.7) 1,144 (65.3) |
127 (29.1) 309 (70.9) |
| BMI, mean (std)*** | 26.4 (4.6) | 28.3 (5.6) |
| Depression, n (%)*** No depression Subsyndromal depression Minor depression Major depression |
1,605 (91.6) 66 (3.8) 50 (2.9) 31 (1.8) |
360 (83.7) 42 (9.8) 17 (4.0) 11 (2.6) |
| Family history of dementia, n (%) Yes No |
278 (21.6) 1,010 (78.4) |
62 (19.3) 259 (80.7) |
| Apolipoprotein ε4 allele, n (%) ε4−/ε4− ε4+/ε4− ε4+/ε4+ |
1,222 (72.7) 429 (25.5) 31 (1.8) |
310 (74.0) 102 (24.3) 7 (1.7) |
*Significance at P < 0.05.
**Significance at P < 0.01.
***Significance at P < 0.001.
Association between opioids and cognitive function
Table 3 presents cognitive function at baseline and the mean change in the cognitive tests from Wave 2 to Wave 3 according to the different levels of exposure to opioids. There was generally a decline in the cognitive function in the study population from baseline to follow-up. However, the severity of decline in the Trail Making B test appeared to increase as the total MED of opioid exposure during the study period increased, from 7.1 in the unexposed group to 13.1 in the group with the highest opioid exposure (MED > 2,940).
Table 3 .
Association between use of opioids (categorized according to quartiles of total MED) and cognitive function (mean and SD)
| MMSEa | Immediate recall | Delayed recall | Digit backb | Spotc | SDMTd | SRTe | CRTf | PPEG (DH)g | PPEG (NDH)h | PPEG (BH)i | Trail Aj | Trail Bk | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| No use Baseline Follow-up Change |
29.2 (1.3) 29.0 (1.3) −0.2 (0.02)*** |
7.0 (2.2) 6.6 (2.2) −0.4 (0.03)*** |
6.1 (2.4) 5.9 (2.3) −0.2 (0.04)*** |
5.2 (2.2) 5.0 (2.2) −0.2 (0.03)*** |
52.9 (5.4) 52.9 (5.4) 0 (0.08) |
50.0 (9.4) 47.3 (9.7) −2.7 (0.1)*** |
278.8 (77.4) 284.7 (68.7) 5.9 (1.1)*** |
329.8 (54.3) 346.2 (58.1) 16.4 (0.8)*** |
13.5 (2.1) 12.2 (2.1) −1.3 (0.03)*** |
12.7 (1.9) 11.6 (2.0) −1.1 (0.03)*** |
10.4 (1.8) 9.4 (1.9) −1.0 (0.03)*** |
34.9 (12.2) 37.3 (14.2) 2.4 (0.2)*** |
81.2 (33.7) 88.3 (37.8) 7.1 (0.5)*** |
| MED 0.01–180.00 Baseline Follow-up Change |
29.2 (1.1) 28.9 (1.3) −0.3 (0.1)*** |
6.6 (2.2) 6.2 (2.3) −0.4 (0.1)*** |
5.7 (2.4) 5.3 (2.5) −0.4 (0.1)** |
4.9 (2.1) 4.8 (2.2) −0.1 (0.1) |
51.2 (4.9) 52.2 (4.9) 1.0 (0.2) |
48.8 (8.6) 46.0 (8.6) −2.8 (0.4)*** |
282.9 (64.5) 285.0 (68.2) 2.1 (3.3) |
332.3 (52.0) 347.4 (59.1) 15.1 (2.8)*** |
13.4 (2.0) 12.2 (2.1) −1.2 (0.1)*** |
12.6 (1.9) 11.6 (2.0) −1.0 (0.1)*** |
10.4 (2.0) 9.4 (2.0) −1.0 (0.1)*** |
36.3 (12.8) 37.8 (12.7) 1.5 (0.6)* |
83.3 (31.2) 91.3 (36.1) 8.0 (1.7)*** |
| MED 180.01–387.30 Baseline Follow-up Change |
28.9 (1.6) 28.9 (1.7) 0 (0.1) |
6.6 (2.3) 6.4 (2.5) −0.2 (0.2) |
5.8 (2.7) 5.8 (2.6) 0 (0.2) |
5.1 (2.0) 4.9 (2.3) −0.2 (0.2) |
51.5 (5.4) 51.2 (5.3) −0.3 (0.4) |
49.5 (10.5) 47.7 (10.5) −1.8 (0.8)* |
277.8 (102.7) 296.8 (108.7) 19.0 (8.3)* |
325.5 (63.5) 350.1 (78.7) 24.6 (5.6)*** |
13.1 (2.4) 12.1 (2.1) −1.0 (0.2)*** |
12.8 (2.5) 11.4 (2.2) −1.4 (0.2)*** |
10.3 (1.9) 9.4 (2.02) −0.9 (0.2)*** |
36.5 (17.0) 38.9 (25.7) 2.4 (1.7) |
83.4 (36.9) 93.8 (39.8) 10.4 (3.0)** |
| MED 387.31–2,940.00 Baseline Follow-up Change |
29.1 (1.2) 28.8 (1.5) −0.3 (0.1)** |
6.9 (2.3) 6.6 (2.4) −0.3 (0.1)* |
6.0 (2.3) 5.7 (2.2) −0.3 (0.1)* |
4.7 (2.0) 4.5 (2.0) −0.2 (0.1) |
51.8 (5.5) 51.7 (5.8) −0.1 (0.3) |
47.7 (9.4) 45.7 (9.3) −2.0 (0.6)*** |
286.9 (91.1) 296.6 (81.9) 9.7 (5.2) |
336.3 (51.9) 354.2 (59.8) 17.9 (3.3)*** |
13.3 (2.1) 11.8 (2.4) −1.5 (0.1)*** |
12.3 (2.2) 11.4 (2.2) −0.9 (0.1)*** |
10.1 (2.0) 9.1 (2.0) −1.0 (0.1)*** |
36.5 (14.5) 39.7 (16.1) 3.2 (0.9)*** |
88.1 (39.4) 95.5 (42.1) 7.4 (2.4)** |
| MED >2,940.00 Baseline Follow-up Change |
29.2 (1.1) 28.8 (1.7) −0.4 (0.9)*** |
6.5 (2.2) 6.3 (2.3) −0.2 (0.1) |
6.0 (2.4) 5.6 (2.3) −0.4 (0.1)** |
4.8 (2.4) 4.5 (2.4) −0.3 (0.1)* |
52.1 (5.6) 52.1 (5.5) 0 (0.3) |
46.8 (9.1) 43.7 (9.8) −3.1 (0.6)*** |
281.4 (69.3) 299.1 (68.4) 17.7 (4.2)*** |
333.3 (47.9) 353.1 (49.8) 19.8 (3.0)*** |
12.8 (2.3) 11.1 (2.4) −1.7 (0.1)*** |
11.9 (2.1) 10.7 (2.3) −1.2 (0.1)*** |
10.0 (2.0) 8.6 (2.2) −1.4 (0.1)*** |
35.0 (9.9) 40.0 (12.4) 5.0 (0.7)*** |
82.6 (31.5) 95.7 (35.6) 13.1 (2.1)*** |
Notes: measures for Trail A, Trail B, SRT and CRT represent response time. Thus, positive values for change indicate cognitive decline. All other measures (MMSE, Immediate Recall, Delayed Recall, Digit Back, Spot, SDMT and PPEG) represent the number of items completed correctly (negative values for change indicate cognitive decline).
aMMSE.
bDigit Span Backwards Test.
cSpot-the-Word Test.
dSymbol Digit Modalities Test.
eSimple Reaction Time.
fChoice Reaction Time.
gPurdue Pegboard Test (Dominant Hand).
hPurdue Pegboard Test (Non-dominant hand).
iPurdue Pegboard Test (Both hands).
jTrail Making Test Part A.
kTrail Making Test Part B.
*Significance at P < 0.05.
**Significance at P < 0.01.
***Significance at P < 0.001.
Table 4 shows the results from the assessment of the association between use of opioids and cognitive function from GLMs. Compared with those not on opioids, individuals exposed to opioids resulting in total MED of greater than 2,940 had significantly lower scores in the MMSE (Model 1: β = −0.34, Model 2: β = −0.35 and Model 3: β = −0.39, P < 0.01). While exposure at this level appeared to affect performance in the Purdue Pegboard Test for dominant hand (Model 1: β = −0.49, Model 2: β = −0.49 and Model 3: β = −0.53), the findings were not statistically significant at P < 0.01.
Table 4 .
Association between use of opioids (categorized according to quartiles of total MED) and cognitive function (β weights and SE)
| MMSEa | Immediate recall | Delayed recall | Digit Backb | Spotc | SDMTd | SRTe | CRTf | PPEG (DH)g | PPEG (NDH)h | PPEG (BH)i | Trail Aj | Trail Bk | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| MED 0.01–180.00 Model 1 Model 2 Model 3 |
−0.09 (0.11) −0.09 (0.11) −0.08 (0.13) |
−0.02 (0.18) −0.03 (0.18) −0.13 (0.21) |
−0.15 (0.18) −0.17 (0.18) −0.24 (0.21) |
0.06 (0.15) 0.06 (0.15) 0.07 (0.18) |
0.38 (0.25) −0.54 (0.50) −0.56 (0.50) |
−0.55 (0.50) −0.56 (0.50) −0.26 (0.60) |
−3.91 (6.30) −3.60 (6.30) −6.64 (7.33) |
−1.26 (3.85) −1.09 (3.86) 1.51 (4.50) |
0.04 (0.18) 0.03 (0.18) 0.05 (0.21) |
0.26 (0.16) 0.24 (0.16) 0.15 (0.19) |
0.05 (0.15) 0.05 (0.15) −0.01 (0.17) |
−0.93 (1.12) −0.85 (1.12) −0.70 (1.33) |
0.84 (2.41) 0.84 (2.41) 3.90 (3.03) |
| MED 180.01–387.30 Model 1 Model 2 Model 3 |
0.13 (0.16) 0.13 (0.16) 0.18 (0.19) |
0.21 (0.27) 0.21 (0.27) 0.45 (0.31) |
0.24 (0.26) 0.23 (0.26) 0.47 (0.31) |
−0.12 (0.24) −0.11 (0.23) −0.09 (0.28) |
−0.26 (0.34) 0.39 (0.77) 0.37 (0.77) |
0.39 (0.77) 0.37 (0.77) 0.41 (0.87) |
13.11 (9.14) 13.72 (9.12) 19.17 (10.65) |
8.27 (6.38) 8.29 (6.40) 12.60 (7.48) |
0.17 (0.25) 0.17 (0.25) −0.02 (0.30) |
−0.17 (0.24) −0.20 (0.24) −0.30 (0.29) |
0.07 (0.22) 0.06 (0.22) 0.02 (0.26) |
−0.06 (1.67) −0.12 (1.66) 0.35 (1.94) |
3.22 (3.68) 2.74 (3.70) 3.65 (4.23) |
| MED 387.31–2,940.00 Model 1 Model 2 Model 3 |
−0.13 (0.13) −0.15 (0.13) −0.17 (0.16) |
0.11 (0.20) 0.08 (0.20) 0.06 (0.25) |
−0.07 (0.20) −0.10 (0.20) 0.002 (0.25) |
−0.09 (0.19) −0.08 (0.19) 0.02 (0.23) |
−0.10 (0.25) 0.25 (0.59) 0.15 (0.59) |
0.25 (0.59) 0.15 (0.59) 0.30 (0.71) |
3.77 (7.09) 4.28 (7.19) −0.28 (8.81) |
1.52 (4.97) 1.83 (5.03) −2.07 (6.27) |
−0.25 (0.20) −0.26 (0.20) −0.20 (0.26) |
0.17 (0.20) 0.17 (0.20) −0.004 (0.25) |
0.06 (0.17) 0.09 (0.17) −0.05 (0.21) |
0.76 (1.31) 1.11 (1.31) 0.44 (1.49) |
0.21 (2.98) −0.004 (2.98) −0.33 (3.93) |
| MED >2,940.00 Model 1 Model 2 Model 3 |
−0.34 (0.13)** −0.35 (0.13)** −0.39 (0.15)** |
0.18 (0.22) 0.16 (0.22) 0.28 (0.25) |
−0.13 (0.22) −0.16 (0.22) 0.01 (0.25) |
−0.19 (0.19) −0.18 (0.19) −0.17 (0.21) |
−0.05 (0.29) −0.87 (0.61) −0.93 (0.61) |
−0.87 (0.61) −0.93 (0.61) −0.37 (0.69) |
11.82 (7.39) 12.52 (7.42) 13.29 (8.35) |
3.34 (4.82) 3.59 (4.84) −1.18 (5.44) |
−0.49 (0.21)* −0.49 (0.09)* −0.53 (0.24)* |
0.02 (0.21) −0.01 (0.21) −0.12 (0.25) |
−0.35 (0.18)* −0.34 (0.18)* −0.43 (0.20)* |
2.56 (1.39) 2.72 (1.42) 2.50 (1.56) |
5.91 (2.89)* 5.58 (2.90) 2.85 (3.48) |
Note: measures for Trail A, Trail B, SRT and CRT represent response time. Thus, positive β values indicate poorer performance relative to no anticholinergic use group. All other measures represent the number of items completed correctly (negative β values indicate poorer performance). Model 1 = unadjusted model; Model 2 = adjusted for age, sex and education; Model 3 = Model 2 + smoking, alcohol consumption, physical activity, stroke, diabetes, hypertension, BMI, depression and family history of dementia.
aMMSE.
bDigit Span Backwards Test.
cSpot-the-Word Test.
dSymbol Digit Modalities Test.
eSimple Reaction Time.
fChoice Reaction Time.
gPurdue Pegboard Test (Dominant Hand).
hPurdue Pegboard Test (Non-dominant hand).
iPurdue Pegboard Test (Both hands).
jTrail Making Test Part A.
kTrail Making Test Part B.
*Significance at P < 0.05.
**Significance at P < 0.01.
***Significance at P < 0.001.
Role of APOE ε-4 allele on the effect of opioids on cognitive function
APOE was significantly associated with performance in Immediate Recall and Delayed Recall. Table 5 presents the effect modification of the association between opioid use and cognitive function by presence of one APOE ε-4 allele (APOE ε-4 +/−) and presence of both alleles (APOE ε-4 +/+). The changes in performance in all the cognitive domains that were assessed were not significantly associated with the interaction term (APOE ε-4 +/−)*opioid. The interaction term (APOE ε-4 +/+)*Ach was significantly associated with decline in Immediate Recall (β = −5.71, P < 0.01) and Delayed Recall (β = −7.38, P < 0.001) test scores among individuals with opioid intake exceeding cumulative total MED of 2,940.
Table 5 .
Effect modification of the association between use of opioids (categorized according to quartiles of total MED) and cognitive function (β weights and SE) by apolipoprotein (APOE) ε4 allele
| MMSEa | Immediate recall | Delayed recall | Digit backb | Spotc | SDMTd | SRTe | CRTf | PPEG (DH)g | PPEG (NDH)h | PPEG (BH)i | Trail Aj | Trail Bk | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| MED 0.01–180.00 +/− +/+ |
−0.12 (0.25) 0.55 (1.09) |
0.37 (0.46) −0.30 (1.87) |
0.16 (0.43) 1.05 (1.61) |
−0.04 (0.38) 0.56 (1.78) |
0.31 (0.53) −1.16 (2.56) |
−1.00 (1.13) 1.91 (5.04) |
11.65 (13.99) −19.59 (55.58) |
−11.08 (9.03) 5.01 (36.04) |
−0.11 (0.39) 1.38 (1.98) |
0.26 (0.36) 0.71 (1.57) |
0.27 (0.32) −0.07 (1.43) |
−0.84 (2.54) −0.37 (11.29) |
−8.72 (6.30) −30.34 (22.74) |
| MED 180.01–387.30 +/− +/+ |
−0.33 (0.39) 0.49 (1.27) |
−0.79 (0.60) −1.10 (2.13) |
−0.68 (0.61) −4.09 (2.13) |
0.08 (0.55) −1.22 (1.86) |
−0.87 (0.82) −0.42 (2.51) |
−0.74 (2.02) 4.21 (6.09) |
−19.57 (22.31) −18.48 (72.85) |
−18.19 (46.99) −18.99 (46.99) |
0.85 (0.58) −0.59 (2.05) |
0.47 (0.57) −3.09 (1.93) |
0.34 (0.52) −2.40 (1.69) |
−0.07 (3.80) −30.47 (13.15) |
−3.08 (8.35) −18.81 (29.03) |
| MED 387.31–2,940.00 +/− +/+ |
−0.004 (0.31) 0.06 (0.77) |
0.02 (0.51) −0.03 (1.27) |
−0.36 (0.48) −1.29 (1.27) |
−0.17 (0.46) −0.56 (1.12) |
−0.25 (0.59) −0.19 (1.51) |
0.23 (1.45) 0.37 (3.70) |
16.32 (16.67) 3.10 (43.70) |
10.46 (12.18) 10.85 (28.38) |
−0.04 (0.53) −1.27 (1.25) |
0.47 (0.49) −0.56 (1.19) |
0.41 (0.40) 0.71 (1.02) |
2.11 (3.05) −7.23 (7.92) |
0.61 (7.98) −13.48 (17.43) |
| MED >2,940.00 +/− +/+ |
−0.07 (0.36) 1.73 (1.26) |
−0.49 (0.53) −5.71 (2.10)** |
−0.55 (0.55) −7.38 (2.12)*** |
0.05 (0.47) −1.16 (1.84) |
0.76 (0.62) −0.75 (2.49) |
−2.29 (1.45) 3.38 (6.02) |
−8.15 (19.06) 39.91 (72.15) |
9.21 (13.84) 59.84 (46.66) |
0.29 (0.50) 0.99 (2.03) |
0.21 (0.51) −3.14 (1.92) |
0.32 (0.46) 2.27 (1.68) |
−1.23 (3.21) 18.61 (13.08) |
7.12 (7.21) 6.83 (8.87)* |
Notes: measures for Trail A, Trail B, SRT and CRT represent response time. Thus, positive β values indicate poorer performance relative to no anticholinergic use group. All other measures represent the number of items completed correctly (negative β values indicate poorer performance). +/− = Presence of one APOE-ε4 allele; +/+ = Presence of two APOE-ε4 allele.
aMMSE.
bDigit Span Backwards Test.
cSpot-the-Word Test.
dSymbol Digit Modalities Test.
eSimple Reaction Time
fChoice Reaction Time.
gPurdue Pegboard Test (Dominant Hand).
hPurdue Pegboard Test (Non-dominant hand).
iPurdue Pegboard Test (Both hands).
jTrail Making Test Part A.
kTrail Making Test Part B.
*Significance at P < 0.05.
**Significance at P < 0.01.
***Significance at P < 0.001.
Discussion
In this study, we observed the effects of opioids on the change in cognitive function within the 4 years from Wave 2 to Wave 3 of the PATH through life study. We found that compared with those not exposed to opioids, individuals with a cumulative total MED of above 2,940 experienced significant decline in their MMSE scores, reflecting decline in global cognitive function. Decline in MMSE scores was seen across the population. However, individuals with cumulative total MED exceeding 2,940 experienced ~0.4 points higher decline compared with those unexposed to opioids. This association remained significant even after adjusting for multiple covariates, indicating that high cumulative exposure to opioid may accelerate the decline of MMSE scores.
Many older adults experience pain daily, with some reporting significant pain that interferes with their normal functioning [34]. The prevalence of chronic pain increases with age [35–37]. This age-related increase is not limited to just joint pain, as commonly expected in the older adult population [38]. The most common causes of chronic pain in this population is arthritis and neuralgias [39]. Compared with those in long-term care facilities, community dwelling older adults are more likely to complain about chronic pain [40, 41]. Chronic pain in older adults should not be ignored as it results in poor daily functioning, affected mood and withdrawal from recreational activities, all of which can adversely affect healthy ageing [42, 43]. With the increase in the prevalence of chronic pain in older adults, the management of pain through opioids has increased [44–46].
Pharmacological intervention for pain in older adults must be approached with caution. The geriatric population is considered a challenging population to treat for pain because of multiple issues related to existing comorbidities, age-related increase in the risk of cognitive impairment and inherent polypharmacy [47]. Several studies have indicated that opioid use does not directly impact cognitive function [48, 49]. However, our study further supports the findings of studies that have shown the detrimental effect of cumulative high doses of opioids on cognition, particularly in older adults [50, 51]. Opioids are associated with episodes of delirium in older adults, which is a risk factor for cognitive impairment [14]. Opioids may also cause hallucinations in older adults [15, 16]. A common side effect of opioids is sedation, which can cause psychomotor impairment [52]. Evidence also indicates that opioids influence learning and memory [53]. Opioids block pain through the disruption of normal neurotransmitter activity in the brain, which could lead to these impairments.
This study has several strengths. The PATH Through Life study used a longitudinal cohort study design on a large sample of population dwelling individuals. The study measured cognitive function using comprehensive neuropsychology assessments and provides us the opportunity to study longitudinal changes in multiple cognitive domains over time. The PATH Through Life study also collected data on multiple covariates, which allowed us to include important covariates in our analysis and further investigate the role of APOE ε-4 genotype as an effect modifier. The PATH study captured the medication use of participants through the PBS medication database. This allowed us to objectively quantify opioid medication use and avoid recall bias by relying solely on self-reported opioid use.
This study also has some limitations. The PBS database captured information of medications that are prescribed and filled. It did not provide information of medication consumption and adherence. We are also unable to obtain information on opioid abuse beyond prescribed opioids. We took a conservative approach in determining exposure to opioids. Individuals with no prescription information during the exposure window were classified as not exposed to opioids. These limitations may have resulted in misclassification of exposed individuals in this study and resulted in underestimation of the measure of effect. Chronic pain and comorbidities associated with it may influence the association between opioid use and cognitive function. These parameters were not available for our analysis. The time period between the Wave 2 and Wave 3 is short, particularly given that cognitive impairment is often insidious with onset occurring far early before symptoms are apparent. The MMSE is susceptible to ceiling and floor effects, and poor sensitivity to change in high scores, which may have affected our study findings. The use of change in cognitive assessment scores is also subject to the effect of regression to the mean.
Our study supports the need to further investigate the role of opioid use on cognitive function in older adults. Future studies with longer follow-up period will allow us to investigate the impact of chronic opioid use. Future studies should also measure days of opioid use, which we were unable to obtain as a secondary measure of opioid exposure. This method will allow us to assess chronic exposure to opioids more effectively.
Conclusion
This study showed that exposure to opioids exceeding a cumulative total MED of 2,940 over a period of 4 years significantly affects global cognition in older adults. Exposure at this level can also potentially affect psychomotor speed. The opioid epidemic is an important public health concern, and this concern should not be disregarded in the older population. Physicians providing care to older adults must be aware of the adverse effects of this option of pain management, particularly in prolonged use of pain medication for chronic pain. Healthcare providers must strive to use alternative options and methods of pain management when feasible, especially in the older population. It is essential for opioid awareness programs to expand to reach older adults as well, as awareness on the adverse effects of opioids on healthy ageing can educate older adults to advocate for themselves when it comes to pain management.
Contributor Information
Malinee Neelamegam, Yale School of Public Health, Yale University, New Haven, CT, USA; College of Public Health, University of South Florida, Tampa, FL, USA.
Janice Zgibor, College of Public Health, University of South Florida, Tampa, FL, USA.
Henian Chen, College of Public Health, University of South Florida, Tampa, FL, USA.
Kathleen O’rourke, College of Public Health, University of South Florida, Tampa, FL, USA.
Chighaf Bakour, College of Public Health, University of South Florida, Tampa, FL, USA.
Lakshminarayan Rajaram, College of Public Health, University of South Florida, Tampa, FL, USA.
Kaarin J Anstey, Centre for Research on Ageing, Health and Wellbeing, Australian National University, Canberra, Australia; School of Psychology, University of New South Wales, Sydney, Australia; Neuroscience Research Australia, Sydney, Australia.
Declaration of Sources of Funding
This work was supported by the National Health and Medical Research Council (NHMRC) (grant numbers 179805 and 1002160 and 418039). M. Neelamegam was funded by the Endeavour Research Fellowship. M. Neelamegam is funded by the National Institutes of Health, Fogarty International Center (NIH, FIC) and the National Institute of Neurological Disorders and Stroke (NINDS) (award number D43TW010540). K.J. Anstey is funded by NHMRC Research Fellowship (grant number 1102694).
Declaration of Conflicts of Interest
None.
References
- 1.Gibson SJ. IASP global year against pain in older persons: highlighting the current status and future perspectives in geriatric pain. Expert Rev Neurother 2007; 7: 627–35. [DOI] [PubMed] [Google Scholar]
- 2.Gagliese L. Pain and aging: the emergence of a new subfield of pain research. J Pain 2009; 10: 343–53. [DOI] [PubMed] [Google Scholar]
- 3.Wilder-Smith OH. Opioid use in the elderly. Eur J Pain 2005; 9: 137–40. [DOI] [PubMed] [Google Scholar]
- 4.Bell JR. Australian trends in opioid prescribing for chronic non-cancer pain, 1986–1996. Med J Aust 1997; 167: 26–9. [DOI] [PubMed] [Google Scholar]
- 5.Boudreau D, von Korff M, Rutter CMet al. . Trends in long-term opioid therapy for chronic non-cancer pain. Pharmacoepidemiol Drug Saf 2009; 18: 1166–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Dunn KM, Saunders KW, Rutter CMet al. . Opioid prescriptions for chronic pain and overdose: a cohort study. Ann Intern Med 2010; 152: 85–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Sandvik R, Selbaek G, Kirkevold O, Husebo BS, Aarsland D. Analgesic prescribing patterns in Norwegian nursing homes from 2000 to 2011: trend analyses of four data samples. Age Ageing 2016; 45: 54–60. [DOI] [PubMed] [Google Scholar]
- 8.Williams RE, Sampson TJ, Kalilani L, Wurzelmann JI, Janning SW. Epidemiology of opioid pharmacy claims in the United States. J Opioid Manag 2008; 4: 145–52. [DOI] [PubMed] [Google Scholar]
- 9.Veal FC, Bereznicki LRE, Thompson AJ, Peterson GM. Use of opioid analgesics in older Australians. Pain Med 2015; 16: 1519–27. [DOI] [PubMed] [Google Scholar]
- 10.Chau DL, Walker V, Pai L, Cho LM. Opiates and elderly: use and side effects. Clin Interv Aging 2008; 3: 273. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Vestergaard P, Rejnmark L, Mosekilde L. Fracture risk associated with the use of morphine and opiates. J Intern Med 2006; 260: 76–87. [DOI] [PubMed] [Google Scholar]
- 12.Ensrud KE, Blackwell T, Mangione CMet al. . Central nervous system active medications and risk for fractures in older women. Arch Intern Med 2003; 163: 949–57. [DOI] [PubMed] [Google Scholar]
- 13.Nurminen J, Puustinen J, Piirtola M, Vahlberg T, Lyles A, Kivela SL. Opioids, antiepileptic and anticholinergic drugs and the risk of fractures in patients 65 years of age and older: a prospective population-based study. Age Ageing 2013; 42: 318–24. [DOI] [PubMed] [Google Scholar]
- 14.Pisani MA, Murphy TE, Araujo KLB, Slattum P, van Ness PH, Inouye SK. Benzodiazepine and opioid use and the duration of ICU delirium in an older population. Crit Care Med 2009; 37: 177. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Lawlor PG. The panorama of opioid-related cognitive dysfunction in patients with cancer. Cancer 2002; 94: 1836–53. [DOI] [PubMed] [Google Scholar]
- 16.Fong HK, Sands LP, Leung JM. The role of postoperative analgesia in delirium and cognitive decline in elderly patients: a systematic review. Anesth Analg 2006; 102: 1255–66. [DOI] [PubMed] [Google Scholar]
- 17.Chahl LA. Opioids-mechanisms of action. Australian Prescriber 1996; 19: 63–5. [Google Scholar]
- 18.Perry VH, Nicoll JA, Holmes C. Microglia in neurodegenerative disease. Nat Rev Neurol 2010; 6: 193–201. [DOI] [PubMed] [Google Scholar]
- 19.Anstey KJ, Christensen H, Butterworth Pet al. . Cohort profile: the PATH through life project. Int J Epidemiol 2011; 41: 951–60. [DOI] [PubMed] [Google Scholar]
- 20.Mellish L, Karanges EA, Litchfield MJet al. . The Australian pharmaceutical benefits scheme data collection: a practical guide for researchers. BMC Res Notes 2015; 8: 634. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Page E, Kemp-Casey A, Korda R, Banks E. Using Australian pharmaceutical benefits scheme data for pharmacoepidemiological research: challenges and approaches. Public Health Res Pract 2015; 25: e2541546. [DOI] [PubMed] [Google Scholar]
- 22.Organization, W.H . The anatomical therapeutic chemical classification system with defined daily doses (ATC/DDD). Oslo: WHO, 2006. [Google Scholar]
- 23.Fine PG, Portenoy RK. A Clinical Guide to Opioid Analgesia. New York, NY: Healthcare Information Programs, 2004. [Google Scholar]
- 24.Vieweg WVR, Lipps WFC, Fernandez A. Opioids and methadone equivalents for clinicians. Prim Care Companion J Clin Psychiatry 2005; 7: 86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Folstein MF, Folstein SE, McHugh PR. "Mini-mental state". A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975; 12: 189–98. [DOI] [PubMed] [Google Scholar]
- 26.Delis DC, Kramer JH, Kaplan E, Thompkins BAO. CVLT: California Verbal Learning Test-Adult Version: Manual. San Antonio, TX: Psychological Corporation, 1987. [Google Scholar]
- 27.Wechsler D. A standardized memory scale for clinical use. J Psychol 1945; 19: 87–95. [Google Scholar]
- 28.Smith A. Symbol Digit Modalities Test. Los Angeles, CA: Western Psychological Services, 1982. [Google Scholar]
- 29.Baddeley A, Emslie H, Nimmo-Smith I. The spot-the-word test: a robust estimate of verbal intelligence based on lexical decision. Br J Clin Psychol 1993; 32: 55–65. [DOI] [PubMed] [Google Scholar]
- 30.Anstey KJ, Dear K, Christensen H, Jorm AF. Biomarkers, health, lifestyle, and demographic variables as correlates of reaction time performance in early, middle, and late adulthood. Q J Exp Psychol A 2005; 58: 5–21. [DOI] [PubMed] [Google Scholar]
- 31.Reitan RM. Validity of the trail making test as an indicator of organic brain damage. Percept Mot Skills 1958; 8: 271–6. [Google Scholar]
- 32.Tiffin J, Asher EJ. The Purdue pegboard; norms and studies of reliability and validity. J Appl Psychol 1948; 32: 234–47. [DOI] [PubMed] [Google Scholar]
- 33.Jorm AF, Mather KA, Butterworth P, Anstey KJ, Christensen H, Easteal S. APOE genotype and cognitive functioning in a large age-stratified population sample. Neuropsychology 2007; 21: 1–8. [DOI] [PubMed] [Google Scholar]
- 34.Gagliese L, Melzack R. Chronic pain in elderly people. Pain 1997; 70: 3–14. [DOI] [PubMed] [Google Scholar]
- 35.Crook J, Rideout E, Browne G. The prevalence of pain complaints in a general population. Pain 1984; 18: 299–314. [DOI] [PubMed] [Google Scholar]
- 36.Badley EM, Tennant A. Changing profile of joint disorders with age: findings from a postal survey of the population of Calderdale, West Yorkshire, United Kingdom. Ann Rheum Dis 1992; 51: 366. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Wolfe F, Ross K, Anderson J, Russell IJ, Hebert L. The prevalence and characteristics of fibromyalgia in the general population. Arthritis Rheum 1995; 38: 19–28. [DOI] [PubMed] [Google Scholar]
- 38.Sternbach RA. Survey of pain in the United States: the Nuprin pain report. Clin J Pain 1986; 2: 49–53. [DOI] [PubMed] [Google Scholar]
- 39.Schnitzer TJ. Update on guidelines for the treatment of chronic musculoskeletal pain. Clin Rheumatol 2006; 25: S22–9. [DOI] [PubMed] [Google Scholar]
- 40.Ferrell BA, Ferrell BR, Osterweil D. Pain in the nursing home. J Am Geriatr Soc 1990; 38: 409–14. [DOI] [PubMed] [Google Scholar]
- 41.Roy R, Thomas M. A survey of chronic pain in an elderly population. Can Fam Physician 1986; 32: 513. [PMC free article] [PubMed] [Google Scholar]
- 42.Won A, Lapane K, Gambassi Get al. . Correlates and management of nonmalignant pain in the nursing home. J Am Geriatr Soc 1999; 47: 936–42. [DOI] [PubMed] [Google Scholar]
- 43.Bernabei R, Gambassi G, Lapane Ket al. . Management of pain in elderly patients with cancer. JAMA 1998; 279: 1877–82. [DOI] [PubMed] [Google Scholar]
- 44.Caudill-Slosberg MA, Schwartz LM, Woloshin S. Office visits and analgesic prescriptions for musculoskeletal pain in US: 1980 vs. 2000. Pain 2004; 109: 514–9. [DOI] [PubMed] [Google Scholar]
- 45.Kelly JP, Cook SF, Kaufman DW, Anderson T, Rosenberg L, Mitchell AA. Prevalence and characteristics of opioid use in the US adult population. Pain 2008; 138: 507–13. [DOI] [PubMed] [Google Scholar]
- 46.Zerzan JT, Morden NE, Soumerai Set al. . Trends and geographic variation of opiate medication use in state Medicaid fee-for-service programs, 1996 to 2002. Med Care 2006; 44: 1005–10. [DOI] [PubMed] [Google Scholar]
- 47.Prommer E, Ficek B. Management of pain in the elderly at the end of life. Drugs Aging 2012; 29: 285–305. [DOI] [PubMed] [Google Scholar]
- 48.Guerriero F, Roberto A, Greco MT, Sgarlata C, Rollone M, Corli O. Long-term efficacy and safety of oxycodone-naloxone prolonged release in geriatric patients with moderate-to-severe chronic noncancer pain: a 52-week open-label extension phase study. Drug Des Devel Ther 2016; 10: 1515–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Raja SN, Haythornthwaite JA, Pappagallo Met al. . Opioids versus antidepressants in postherpetic neuralgia: a randomized, placebo-controlled trial. Neurology 2002; 59: 1015–21. [DOI] [PubMed] [Google Scholar]
- 50.Puustinen J, Nurminen J, Löppönen Met al. . Use of CNS medications and cognitive decline in the aged: a longitudinal population-based study. BMC Geriatr 2011; 11: 70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51.Dublin S, Walker RL, Gray SLet al. . Prescription opioids and risk of dementia or cognitive decline: a prospective cohort study. J Am Geriatr Soc 2015; 63: 1519–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52.Schor JD, Levkoff SE, Lipsitz LAet al. . Risk factors for delirium in hospitalized elderly. JAMA 1992; 267: 827–31. [PubMed] [Google Scholar]
- 53.McGaugh J, Introini-Collison I, Castellano C. Involvement of opioid peptides in learning and memory. In: Opioids II. Berlin: Springer, 1993; 429–47. [Google Scholar]