Centralized Pain Diagnoses in Older Adults With and Without Cancer

Patients in this cross-sectional analysis were at least 65 years of age with at least one outpatient visit between November 1, 2015 and December 23, 2021 within our healthcare system were included. For patients with more than one outpatient visit during the study period, only the first visit was included. The primary outcome was the proportion of older adults, with and without cancer, who had at least one diagnosis of a COPC coded in their record. Secondary outcomes included the rate of COPCs, associations of comorbid conditions (via Elixhauser) or demographic variables with a COPC diagnosis, and age given the known association with chronic pain ( 16 ). The Elixhauser Comorbidity Index is a risk adjustment tool that quantifies comorbidity burden using distinct comorbidity categories in administrative data with strong performance ( 16 ). The electronic medical records were queried using discrete variable search and text search tools ( 17 ). Pertinent demographic and clinical data were abstracted using the tools. Demographic data included age, race/ethnicity, gender, and outpatient visit dates. Clinical data included a set of COPC diagnoses identified via a validated ICD10 code set ( 14 ), Elixhauser comorbidities, and history of cancer (lymphoma, metastatic cancer, or solid tumor without metastases). The COPC diagnoses of interest included: fibromyalgia, irritable bowel syndrome, interstitial cystitis/bladder pain syndrome (IC/BPS), chronic prostatitis (CP), vulvodynia (VVD), migraine headaches (MH), chronic tension type headaches (cTTH), temporomandibular joint dysfunction (TMD), chronic low back pain, endometriosis (Endo), myalgic encephalomyelitis/chronic fatigue syndrome (CFS/ME) ( 14 , 15 ). This study was exempted from the University of Michigan institutional review board per exemption 4 (ii) at 45 Code of Federal Regulations 46.104(d). Descriptive statistics were used to summarize the demographic variables of the sample population. Categorical variables were summarized with counts and percentages, and numeric variables were summarized with medians and interquartile ranges to account for skewness in their distributions. Missing values were retained in the data, and only the spline regressions had exclusions due to listwise deletion. This was done to represent as many people in the analyses as possible. We compared the COPC and non-COPC groups’ baseline variables using univariable logistic regressions. After comparing COPC and non-COPC groups, we continued the analysis with just the COPC group. We described the connection between COPC diagnoses and cancers with univariable regressions. For this we used logistic regressions when looking at the presence of cancer, and negative binomial regressions when looking at the counts of COPC diagnoses. The negative binomial regressions were implemented over Poisson regressions so as to account for overdispersion in count data. We also used cubic spline regressions, with four knots, to determine the variable impact of age on the presence of COPC diagnoses. In addition to age, we adjusted for Elixhauser comorbidity, Hispanicity, race, and gender. Because of the many hypothesis tests that we conducted (there were 105), we adjusted our p -values with the Benjamini-Hochberg correction, thereby keeping our false discovery rate at 0.05. The 95% confidence intervals were specified with the Benjamini–Hochberg adjustment as well.

There were 320,727 older adults that met study criteria and were included ( Table 1 ). Of these, 15.5% ( n = 49,750) carried at least 1 COPC diagnosis. Of those with COPCs, 62.3% ( n = 30,979) were female and the median age was 67 years (IQR 65–73). Of those without COPCs (84.5%, n = 270,977), 49.6% ( n = 134,453) were female and the median age was 69 years (IQR 66–75). Compared to the non-COPC group, those with COPCs were significantly more likely to be female and younger. Elixhauser Comorbidity Index was significantly higher in the COPC compared to the non-COPC group (2.0 [IQR 1.0–5.0] vs 1.0 [IQR 0–3.0], p = <0.001; specific Elixhauser comorbidities in Supplementary Table I ). Hispanicity and race distribution was not statistically significant between groups, except that Black Americans comprised 6.1% of the COPC group versus 6.8% of the non-COPC group (OR 0.89, 95% CI [0.856, 0.935], p = <0.001). Older adults with COPCs were more likely to carry any cancer diagnosis compared to those without: 13,136 (26.4%) vs. 69,298 (25.6%), OR 1.04, p = <0.001 ( Table 1 ). The incidence rate of COPCs was 1.033 times higher in older adults with cancer compared to older adults without cancer (95% CI [1.014, 1.053], p = 0.001; Table 2 ). When stratifying by cancer type (lymphoma, metastatic cancer, and solid tumor without metastasis), the rate of COPCs remained significantly higher in those with cancer compared to those without. The rate of COPCs was 1.063 (95% CI [1.016, 1.114], p = 0.008), 1.045 (95% CI [1.023, 1.068], p = <0.001), and 1.023 (95% CI [1.001, 1.045], p = 0.039) times higher for those diagnosed with lymphoma, metastatic cancer, and solid tumor without metastasis, respectively, compared to those without. The rate of COPCs was also higher in females (IRR 1.11, 95% CI [1.090, 1.129], p = <0.001) and decreased with age (IRR 0.996, 95% CI [0.995, 0.997], p = <0.001). When stratifying by COPC type, older adults with any cancer diagnosis were significantly more likely to carry a diagnosis of TMD [110 (34.2%) vs. 13026 (26.4%), OR 1.45, p = 0.003], cLBP [8072 (26.9%) vs. 5064 (25.6%), OR 1.069, p = 0.002], and CFS/ME [1455 (31.1%) vs. 11681 (25.9%), OR 1.294, p = <0.001] ( Table 3 ). Older adults with any cancer diagnosis were significantly less likely to carry a diagnosis of FM [1280 (25.0%) vs. 11856 (26.6%), OR 0.921, p = 0.024]. Older adults with solid tumors without metastasis also were more likely to carry diagnoses of cLBP [5647 (18.8%) vs. 3431 (17.4%), OR 1.104, p = <0.001] and CFS/ME [1039 (22.2%) vs. 8039 (17.8%), OR 1.318, p = <0.001] and were less likely to carry a FM diagnosis [826 (16.1%) vs 8252 (18.5%), OR 0.848, p = <0.001]. cTTH diagnoses were also more likely in this group [105 (23.0%) vs. 8973 (18.2%), OR 1.34, p = 0.014]. Older adults with metastatic cancer were more likely to carry diagnoses of MH [2080 (18.5%) vs. 6696 (17.4%), OR 1.075, p = 0.015], cTTH [99 (21.7%) vs. 8677 (17.6%), OR 1.294, p = 0.036], TMD [80 (24.8%) vs. 8696 (17.6%), OR 1.548, p = 0.001], cLBP [5440 (18.1%) vs. 3336 (16.9%), OR 1.091, p = 0.001], and CFS/ME [994 (21.3%) vs. 7782 (17.3%), OR 1.296, p = <0.001]. IC/BPS diagnoses were less likely in this group [138 (14.4%) vs. 8638 (17.7%), OR 0.779, p = 0.012]. Older adults with lymphoma were more likely to carry diagnoses of cLBP [1028 (3.4%) vs. 574 (2.9%), OR 1.187, p = 0.002] and CFS/ME [217 (4.6%) vs. 1385 (3.1%), OR 1.537, p = <0.001]. All relationships between age and likelihood of COPCs were statistically significant except for cTTH ( p = 0.054; Figure 1 ). Four main patterns emerged in the relationships between age and likelihood of COPCs in older adult cancer survivors. The first of these was a peak in likelihood in the youngest old age group (65–74 years) followed by a decrease thereafter. FM, IBS, and IC/BPS followed this pattern. cTTH followed this pattern but was not statistically significant. The second pattern was a peak in likelihood in the youngest old and a second peak in the oldest old (≥ 85 years). VVD, CP, and TMD followed this pattern. The third pattern was a steady decline in likelihood of COPCs as age increased. MH and Endo followed this pattern. Finally, the fourth pattern was a steady increase in likelihood of COPCs as age increased. cLBP and CFS/ME followed this pattern.

The face of cancer survivorship is changing and a growing number of older adults are living with cancer. By some estimates, older adults will account for 73% of cancer survivors by 2040 ( 9 ). Despite this, literature focusing on their unique experience of cancer survivorship is sparse compared to their younger counterparts ( 56 ). For the first time, this study shows that older adult cancer survivors are more likely to carry a COPC diagnosis than their peers without cancer. By definition, COPCs have underlying nociplastic etiology, which requires alternative management than other types of pain that can be seen in cancer survivorship. This study highlights the need for greater awareness of the risk of COPCs when treating pain and other symptoms in older adult cancer survivors. Additionally, we found a variable relationship between specific COPCs and age. While the prevalence of most COPCs seems to decrease with age, our results were notable for a peak in some COPCs (VVD, CP, and TMD) in the oldest old group of cancer survivors, which is incongruent with existing literature in the general population. This suggests that the oldest old adult cancer survivors are subject to a greater burden of these COPCs compared to their younger peers, both with cancer and without. Further epidemiological studies should be undertaken to evaluate racial disparities and the temporal association between COPC and cancer diagnoses in older adult cancer survivors.

Although the prevalence of chronic pain in older adult cancer survivors has been described, this study is the first to our knowledge to stratify by pain etiology (i.e., presence of a COPC indicating nociplastic pain etiology) and to examine changes in prevalence of COPCs in older adulthood and cancer survivorship in granular detail. Our study found that older adults carrying at least one COPC diagnosis were statistically more likely to have any cancer present and that all types of cancer (i.e., lymphoma, metastatic cancer, and solid tumor without metastasis) were associated with an increased incidence in the number of COPCs. The increased likelihood of COPC diagnoses in older adults with cancer compared to those without is consistent with findings of increased mortality from cancer in adults with chronic pain and a nearly two-fold increase in the prevalence of chronic pain in cancer survivors compared to the general population ( 18 , 19 ). It is generally accepted that cancer is associated with numerous painful episodes related to the disease itself or its treatment, but whether or not cancer survivors are more likely to have chronic pain prior to diagnosis or develop it after being diagnosed is less clear ( 20 ). Estimates for the prevalence of chronic pain in older adults vary widely, with some studies estimating the prevalence at upwards of 70% ( 2 ). Interestingly, our study found that approximately one in six older adults, or only 15.5%, carry a diagnosis for a COPC. There could be multiple reasons for this difference: that symptoms of COPCs are misattributed to usual processes of aging or other comorbidities, that some COPCs are more often mediated by “top-down” processes that are heritable and onset at a younger age, or that chronic pain fluctuates over time rather than accumulating ( 21 , 22 ). Additionally, COPCs do not encompass the entire range of chronic pain conditions, e.g., osteoarthritis, and so the prevalence of COPCs may accurately reflect the distribution of different chronic pain syndromes in this population. Older adults with a COPC diagnosis were significantly more likely to be female, younger, and have a higher Elixhauser Comorbidity Index than their counterparts who did not carry a COPC diagnosis. This is consistent with literature; in multiple large population studies investigating fibromyalgia, approximately 60% of the sample were female, though some studies place these estimates much higher at 80–90% ( 23 – 26 ). Literature describing gender differences for other COPCs reflects at least a slight female preponderance ( 27 ). Data on the relationship between age and presence of COPCs demonstrates a non-linear relationship, with numerous studies showing the prevalence of many COPCs peaking at middle age and decreasing thereafter ( 21 ). Our study supports this trend when considering COPCs in aggregate, with the median age in the COPC group being 67 years compared to 69 years in the non-COPC group. Strikingly, our results showed that significantly fewer Black Americans were diagnosed with a COPC, comprising only 6.1% of the COPC group compared to 6.8% of the non-COPC group. Literature has clearly demonstrated that Black American adults experience more severe pain, greater levels of pain-related disability, and disparities in its treatment ( 28 – 30 ). However, previous studies investigating COPC prevalence have reported mixed results based on race or ethnicity ( 31 , 32 ). Causality of these disparities need further research, particularly given that chronic widespread pain and COPCs have been linked to negative outcomes in aging like worsening frailty or increasing disability, and development of other debilitating conditions like long COVID, this finding in a population of older adult cancer survivors deserves special attention ( 32 – 35 ). Similar to the general sample, COPCs were more likely to be present in female cancer survivors than their male counterparts. Notably, when stratifying for COPC type, older adults with FM were less likely to carry a diagnosis for any cancer compared to those without FM. This relationship between cancer diagnosis and FM was only significant when examining cancer diagnoses in aggregate, but not when examining just those with metastatic cancer or lymphoma, where the odds of cancer did not differ between those with and without FM. Multiple studies have enumerated the challenges in evaluating the mortality risk associated with fibromyalgia, including changes in diagnostic criteria over time and underdiagnosis ( 36 , 37 ). There could also be diagnostic overshadowing with widespread pain being attributed to cancer, rather than FM, which would lead to less diagnostic coding of FM. Some have even demonstrated a decreased cancer mortality rate associated with fibromyalgia, with the authors surmising that increased utilization of healthcare services allows for early detection of cancer ( 38 ). The evidence has been clearer that adults with chronic widespread pain characteristic of fibromyalgia experience earlier and excess all-cause mortality ( 39 , 40 ). In our population, those with a COPC diagnosis had higher Elixhauser score, indicating a higher risk of mortality. Thus, it is possible that the lower odds of FM diagnosis in older adult cancer survivors with any cancer present is a result of survivor bias, which increases with age in general. Unlike FM, older adults carrying a cLBP and CFS/ME diagnosis were more likely to be diagnosed with any cancer type, which remained significant when stratifying by cancer type (i.e., lymphoma, metastatic cancer, or solid tumor without metastases). Literature describing the prevalence of cLBP in cancer survivors pre- and post-cancer diagnosis is lacking, but CFS/ME has been associated with an increased risk of subsequently developing non-Hodgkin lymphoma but not a statistically significant increase in cancer-related mortality ( 41 , 42 ). While the current study does not elucidate whether or not cLBP and CFS/ME diagnoses preceded cancer diagnoses, the elevated prevalence of these syndromes in all types of cancer is notable because of their overlap in presentation with common cancer related symptoms. Given the need for different pain management strategies based on etiology, this finding underscores the importance of being aware of the potential for COPCs in this population. The other notable finding when examining differences in COPC rates between different cancer types is that older adults with MH, cTTH, TMD, cLBP, and CFS/ME diagnoses had greater odds of having metastatic cancer present. Given the recency of COPCs as a diagnostic entity and methodology to identify them in administrative datasets, the dearth of literature describing this is unsurprising. This finding suggests a disproportionately higher burden of COPCs in older adults with metastatic cancer, who already have higher symptom burden, compared to their peers without metastases. When examining the relationships between age and likelihood of specific COPCs in older adult cancer survivors four patterns emerged. The first of these was a peak in early older adulthood, which included FM, IBS, and IC/BPS. This is consistent with the literature describing trends in the general population ( 21 ). The second pattern in which the likelihood of the COPC peaked in early older adulthood and in the oldest old, included VVD, CP, and TMD. There is no literature at present to describe the prevalence of VVD in older adult cancer survivors, however literature in the general population suggests a drop-off in VVD diagnosis after age 70 ( 43 ). Other studies have shown that VVD trajectories vary according to age at first pain onset and some chemotherapies are known to be associated with development of VVD, which may explain the initial peak in early adulthood and secondary peak in the oldest old group of the study population ( 44 , 45 ). CP also followed a pattern of peaking in early adulthood and in the oldest-old. Like VVD, the literature demonstrates a greater prevalence in middle age, with decreases thereafter ( 46 ). Interestingly, CP has been shown to be associated with an increased odds for development of prostate cancer and colorectal cancer, which could account for the differing trajectory in our study population ( 47 , 48 ). Finally, the prevalence of TMD in the general population has been reported to decrease after age 70 ( 49 ). Like VVD, some cancer treatments, e.g. radiotherapy, have been associated with TMD symptomatology and may account for the peak in the oldest old of the study population ( 50 ). The third pattern in which the likelihood of the COPC steadily declined with age included MH and Endo. This is consistent with the literature describing trends in the general population, though it should be noted that endometriosis is not exclusively a premenopausal disease as is sometimes thought ( 51 , 52 ). Finally, the fourth pattern in which the likelihood of the COPC steadily increased with age included cLBP and CFS/ME. This is consistent with the literature describing trends in the general population, though historically research into cLBP has focused on working age adults. Research in cLBP in older adulthood has found a much higher prevalence, up to 36%, compared to younger adults with estimates ranging from 4.2–19.6% ( 53 ). Literature describing the prevalence of CFS/ME in older adulthood is sparse, with most studies finding a peak in middle age ( 54 , 55 ). This study has limitations that are important to acknowledge. As noted at multiple points in the discussion, a better understanding of the temporal trends of developing cancer and COPCs is important to understand the relationship between the two, which is not addressed by the current study. While use of ICD10 codes is invaluable in identifying COPCs in large swaths of the population, ICD coding is inherently limited in its ability to convey clinical complexity. We have attempted to address this by adjusting for medical comorbidity using validated measures including Elixhauser comorbidity. Additionally, our study is subject to limitations including missing information in the EHR. All data for this study was extracted from a database comprising EMRs from a large, academic medical center in the United States, which is not necessarily representative at a national level. We have adjusted additionally for Hispanicity, race, and gender to mitigate some of these limitations. While large-scale, prospective, epidemiological studies should be completed in the future, our current methodology allows for preliminary elucidation of these trends.

Sources show 100 million Americans suffer from chronic pain ( 1 ). The prevalence of chronic pain is known to increase with age, being reported in approximately 60–73% of patients over the age of 65 in developed countries ( 2 ). It is clear that age is associated with chronic pain, but its relationship to nociplastic pain is unknown ( 3 ). Pain can be characterized as nociceptive, neuropathic, or nociplastic. Nociplastic pain is caused by dysfunctional alterations in nociception mediated by the central nervous system in the absence of actual tissue damage, lesions, or disease of the somatosensory nervous system ( 4 , 5 ). Nociplastic pain conditions can occur due to central sensitization or central hypersensitivity ( 6 ). Central sensitization is, in part, caused by ongoing peripheral nociceptive input, as seen in patients with cancer-related pain ( 6 ). As with older adult populations, literature examining nociplastic pain in cancer survivors is sparse. Gaps in understanding nociplastic pain in older adults and cancer survivors place these already vulnerable populations at even greater risk: opioids, the cornerstone of cancer pain management, are fraught with risk when used in older adults and are associated with worsening nociplastic pain conditions when used chronically ( 7 , 8 ). By next year, it is anticipated that there will be more than 13 million cancer survivors over the age of 65, nearly 40% of whom will suffer chronic pain of undefined etiology ( 9 – 13 ). Given the association of age and chronic pain, the risks of pain medication in older adults, and the large population of older adults at risk of central sensitization as described here, it is clear there is a need to study the risks of nociplastic pain in older adults with and without cancer. Chronic overlapping pain conditions (COPCs) are a set of nociplastic pain conditions that have been identified in the literature ( 5 ). Diagnoses that are considered COPCs include fibromyalgia (FM), chronic low back pain (cLBP), irritable bowel syndrome (IBS), and chronic migraine headaches (MH) ( 5 ). Owing to their seemingly disparate presentations, these diagnoses have largely been studied in silos, so their true impact on older adults is unknown ( 5 ). Recently, studies have been conducted that have validated ICD10 and ICD9 codes to study COPCs in administrative data ( 14 , 15 ). These studies have identified the diagnostic codes used in medical records that accurately identify COPCs, which represent a subset of nociplastic pain conditions. By utilizing COPC diagnostic codes, we are able to study nociplasticity across populations, including older adults with and without cancer. Therefore, the objective of this study was to utilize administrative data to determine the impact of COPCs on older adults with and without cancer to lay the groundwork for future work examining medication use trends in these populations.

Supplemental data for this article can be accessed online at https://doi.org/10.1080/15360288.2026.2701348 .