Abstract
Introduction:
Glucosamine and chondroitin are supplements that are often, but not always, used in combination for arthritis and joint pain. Multiple studies have suggested that glucosamine and chondroitin may be associated with reduced risk of several diseases, as well as all-cause, cancer- and respiratory disease-specific mortality.
Methods:
Nationally representative data from the National Health and Nutrition Examination Survey (NHANES) were used to further evaluate the association between glucosamine and chondroitin with mortality. Participants include 38,021 adults, ages 20+ years and older, who completed the detailed NHANES between 1999 and 2014. Participants were followed for death through linkage with the National Death Index through the end of 2015, over which time 4905 deaths occurred. Adjusted hazard ratios (HRs) for overall and cause-specific mortality were estimated using Cox regression models.
Results:
Despite glucosamine and chondroitin use appearing to be inversely associated with mortality in the minimally adjusted models, no association was observed in multivariable models (glucosamine: HR = 1.02; 95% confidence interval [CI]: 0.86–1.21, chondroitin: HR = 1.04, 95% CI: 0.87–1.25). No association with cancer mortality or other mortality rate was observed after multivariable adjustment. There was a suggestive, nonsignificant inverse association for cardiovascular-specific mortality (glucosamine HR = 0.72; 95% CI: 0.46–1.15, chondroitin: HR = 0.76; 95% CI: 0.47–1.21).
Conclusion:
The lack of significant relationship between glucosamine and chondroitin use and all-cause or cause-specific mortality after adjusting extensively for multiple covariates in this nationally representative adult population was in contrast to prior literature. Given the limited power to explore the cause-specific mortality, future well-powered studies will be needed to better understand the potential association with cardiovascular-specific mortality.
Keywords: glucosamine, chondroitin, mortality, cohort
Introduction
Glucosamine and chondroitin are dietary supplements that are available as over-the-counter products in the United States. These supplements are commonly used by older adults1,2 to treat joint pain and arthritis, often as alternatives to nonsteroidal anti-inflammatory drugs (NSAIDs). Glucosamine and chondroitin have a favorable side effect profile,3–5 compared with NSAIDs and equivalent drugs.6–11 Glucosamine and chondroitin are often, but not always, combined (and are sometimes further combined with methylsulfonylmethane [MSM] in joint health supplements). Although often used for joint pain, existing studies show varied results of the effectiveness of glucosamine and chondroitin in relieving joint symptoms.12–16
Emerging evidence suggests that glucosamine may have a role in reducing risk of colorectal and lung cancer, and mortality from cancer, cardiovascular disease, and respiratory disease compared with those not taking glucosamine supplements.17–28 In the VITamins And Lifestyle (VITAL) study, a cohort of 77,510 persons followed for a mean of 6.8 years over which time 5362 deaths occurred, the adjusted hazard ratio (HR) associated with current use of glucosamine was 0.82 (95% confidence interval [CI]: 0.75–0.9), and 0.86 (95% CI: 0.78–0.96) for current use of chondroitin.19 The study found significant risk reductions in mortality due to cancer and other causes of death for those taking glucosamine/chondroitin compared with those who did not.
If glucosamine/chondroitin use reduces the risk of overall mortality, it may be due to their role in inflammation. In vitro and animal research has shown promising evidence of this.29–34 Existing human studies (both observational and clinical trials), show the reduction of some inflammatory biomarkers, including high-sensitivity C-reactive protein (hsCRP) and prostaglandin E2(PGE2).35–38 hsCRP is correlated with both the all-cause and cancer-related mortality,39 alongside other markers of inflammation.40
A recently published study looked at the impact of glucosamine/chondroitin on mortality using the data from the National Health and Nutrition Examination Survey (NHANES).22 While this study observed glucosamine/chondroitin use to be significantly associated with reduced mortality, power and adjustment for covariates were limited. Therefore, this study will expand on the results using NHANES data by examining the association between glucosamine/chondroitin and both the all-cause and cause-specific mortality in a larger study with a broader range of covariates. Given some evidence of heterogeneity by follow-up time and calendar time25,41 in prior analyses of glucosamine/chondroitin (in the context of inflammation and cancer), this study will conduct stratified analyses by these two time constructs.
Methods
Population
The study used data from NHANES, a nationally representative survey that assesses the health and nutritional status of civilian, noninstitutionalized adults and children in the United States. NHANES collected information on the demographic characteristics of the study population, as well lifestyle factors, health history, use of prescription medications, dietary supplements, and diet.
This analysis used data from NHANES from 1999 to 2014. Of the persons identified within the NHANES who were ages 20+ and who participated in surveys between 1999 and 2014 (N = 41,659), participants with missing information on any supplement use (N = 42), those who were pregnant (N = 1344), those with a history of rheumatoid arthritis (N = 2194), and those who were ineligible for mortality follow-up (n = 58) were excluded, resulting in a final sample size of 38,021 persons.
Participants in the study provided written informed consent. Study procedures were approved by the NCHS Research Ethics Review Board. The Memorial Sloan Kettering Cancer Center Institutional Review Board concluded that the analysis did not constitute human subjects research and subsequently did not require human subjects' approval.
Exposure
The use of glucosamine/chondroitin was assessed through interview and inventory. Participants reported supplement use in the prior 30 days, including any use and number of days of use. They provided physical evidence of each supplement, or if unavailable, recalled each product taken. Supplement products included could be capsules, tablets, et cetera, and be produced as single or combined products. Then, the ingredients' database from NHANES was used to identify specific supplements containing either glucosamine or chondroitin as an ingredient, which was used to define exposure variables.
There were two primary exposures in this study: regular use of glucosamine and regular use of chondroitin (defined by use on 20 or more of the prior 30 days for both), compared with those who report neither. Regular users were defined as such, so that persons in the exposed group would have used the exposure frequently enough to have a potential biologic effect.17 Those who used glucosamine or chondroitin for <20 days in the prior 30 days were classified with nonusers as no regular glucosamine use.
The secondary exposures examined were associations for glucosamine/chondroitin, defined as a three-level variable: regular use of glucosamine/chondroitin, regular use of glucosamine alone, or regular use of neither glucosamine nor chondroitin. Associations for the combined products were examined, as they are commonly prepared and used in combination, although sometimes glucosamine alone is used (it was not possible to examine chondroitin alone as it is rarely used in the absence of glucosamine). MSM was also examined in secondary analyses due to the frequency with which it is combined with glucosamine/chondroitin supplements.
Outcome
The primary outcome in this study was the all-cause mortality, as defined by any deaths occurring from the time of baseline questionnaire through the end of 2015. All deaths were identified by linkage with the National Death Index (NDI) data.42
In secondary analyses, cause of death was classified based on the International Classification of Diseases (ICD-10th revision) as cardiovascular-, cancer-, or other mortality.
Statistical analyses
Survival analyses were conducted using Cox proportional hazard models to estimate HRs and the corresponding 95% CIs, comparing users with nonusers. Both minimally adjusted models adjusting for age and sex, and multivariable models, were conducted, both with age as the time-axis of analysis. Covariates in multivariable models were selected a priori to include factors associated with glucosamine use and/or the overall or disease-specific mortality.
Multivariable models adjust for age and sex as well as race/ethnicity (non-Hispanic white; non-Hispanic black; Mexican American; other Hispanic; other race/mixed-race), educational status (<high school; high school graduate or equivalent; some college or associate degree; college graduate or above), marital status (married/living with partner; widowed; divorced/separated; never married), poverty–income ratio (<1 [lowest income]; 1 to <2; 2 to <4; >4 [highest income]), and health insurance (insured; noninsured). Lifestyle factors, including body mass index (BMI; <18.5; 18.5 to 25, 25 to 27.5; 27.5 to <30; >30), alcohol use (<1 drink/month; >1 drink/month to 4 drinks/week; >4 drinks/week to 2 drinks/day; >2 drinks/day), smoking (never; former; current with pack-year data computed from the survey), medication use (aspirin: never; low; high; nonaspirin NSAID use: never; low; high), and physical activity (none; moderate; vigorous), were also incorporated.
In addition, self-reported health status (excellent/very good; good/fair; poor), a morbidity score (none; low; high), and history of arthritis (no; yes) were included. The morbidity score was calculated using beta-coefficients of age-adjusted, sex-specific proportional hazard models of death based on models with specific health conditions (listed in the footnote of Table 1). A risk score per participant was created, using the natural log of the coefficients for the HR for death based on the subject's own group of health conditions compared with a subject with no conditions. A missing indicator was added for each variable where data were unavailable to maximize the number of individuals in the analysis. As aspirin and non-aspirin NSAID use were only available from 1999 to 2004, sensitivity analyses were restricted to these cycles, with and without adjustment for both.
Table 1.
Population Characteristics, by Glucosamine Use
| Population characteristics | Cohort, N (weighted %), 38,021 (100) | Regulara glucosamine users, N (weighted %), 1181 (3.5) | Not regular glucosamine users, N (weighted %), 36,816 (96.5) |
|---|---|---|---|
| Sociodemographic | |||
| Age | |||
| 20–29 | 6510 (18.9) | 28 (0.5) | 6480 (99.5) |
| 30–39 | 6503 (19.1) | 48 (0.9) | 6454 (99.1) |
| 40–49 | 6774 (20.9) | 111 (2.1) | 6660 (97.9) |
| 50–59 | 5587 (17.5) | 222 (5.3) | 5362 (94.7) |
| 60–69 | 5911 (11.8) | 341 (8.2) | 5565 (91.8) |
| 70–79 | 4036 (7.5) | 268 (8.2) | 3759 (91.8) |
| 80+ | 2700 (4.2) | 163 (6.3) | 2563 (93.7) |
| Sex | |||
| Male | 19,104 (49.2) | 558 (3.3) | 18,535 (96.7) |
| Female | 18,917 (50.8) | 623 (3.6) | 18,281 (96.4) |
| Race/ethnicity | |||
| Non-Hispanic white | 17,818 (69.6) | 840 (4.3) | 16,963 (95.7) |
| Non-Hispanic black | 7789 (10.9) | 73 (0.7) | 7713 (99.3) |
| Mexican American | 6784 (8.0) | 95 (1.1) | 6684 (98.9) |
| Other Hispanic | 2847 (5.4) | 72 (1.8) | 2774 (98.2) |
| Other race/mixed race | 2783 (6.1) | 101 (3.3) | 2682 (96.7) |
| Missing | |||
| Educational status | |||
| <High school | 10,684 (18.4) | 171 (1.6) | 10,509 (98.4) |
| High school graduate/GED or equivalent | 8853 (24.0) | 239 (2.8) | 8605 (97.2) |
| Some college or associate degree | 10,457 (30.5) | 362 (3.6) | 10,088 (96.4) |
| College graduate or above | 7974 (27.0) | 409 (5.1) | 7561 (94.9) |
| Missing | 53 (0.1) | 0 (0.0) | 53 (100.0) |
| Marital status | |||
| Married/living with partner | 22,515 (62.6) | 839 (4.3) | 21,667 (95.7) |
| Widowed | 3285 (5.8) | 147 (5.1) | 3133 (94.9) |
| Divorced/separated | 5032 (12.3) | 124 (2.7) | 4903 (97.3) |
| Never married | 6766 (18.0) | 65 (0.9) | 6697 (99.1) |
| Missing | 423 (1.3) | 6 (1.2) | 416 (98.8) |
| Poverty-to-income ratio | |||
| <1 (Lowest income) | 7050 (13.1) | 83 (1.1) | 6964 (98.9) |
| 1 to <2 | 9138 (19.1) | 220 (2.5) | 8913 (97.5) |
| 2 to <4 | 9432 (27.3) | 309 (3.2) | 9118 (96.8) |
| ≥4 (Highest income) | 9212 (33.7) | 490 (5.3) | 8715 (94.7) |
| Missing | 3189 (6.9) | 79 (3.0) | 3106 (97.0) |
| Health insurance | |||
| Insured | 29,234 (80.5) | 1076 (4.1) | 28,137 (95.9) |
| not insured | 8602 (19.0) | 105 (1.1) | 8494 (98.9) |
| Missing | 185 (0.5) | 0 (0.0) | 185 (100.0) |
| Lifestyle factors | |||
| Body mass index (kg/m2) | |||
| Underweight (<18.5) | 631 (1.7) | 10 (1.8) | 621 (97.2) |
| Normal weight (18.5 to <25) | 10,922 (30.5) | 303 (3.0) | 10,610 (97.0) |
| Lower overweight (25 to <27.5) | 6759 (17.8) | 241 (3.8) | 6514 (96.2) |
| Upper overweight (27.5 to <30) | 6047 (15.6) | 211 (4.0) | 5832 (96.0) |
| Obese (≥30) | 12,861 (32.7) | 401 (3.6) | 12,453 (96.4) |
| Missing | 801 (1.7) | 15 (2.3) | 786 (96.4) |
| Alcohol use | |||
| <1 Drink/month | 16,673 (38.5) | 530 (3.5) | 16,131 (96.5) |
| ≥1 Drink/month to <4 drinks/week | 9832 (28.8) | 320 (3.6) | 9504 (96.4) |
| ≥4 Drinks/week to <2 drinks/day | 6002 (18.6) | 223 (4.0) | 5778 (96.0) |
| ≥2 Drinks/day | 1887 (5.6) | 47 (2.8) | 1839 (97.2) |
| Missing | 3627 (8.4) | 61 (2.2) | 3564 (97.8) |
| Smoking | |||
| Never | 20,315 (53.2) | 650 (3.7) | 19,658 (96.3) |
| Former (pack-years Tertile 1) | 2580 (6.9) | 113 (4.6) | 2465 (95.4) |
| Former (pack-years Tertile 2) | 2625 (6.9) | 125 (5.4) | 2496 (94.6) |
| Former (pack-years Tertile 3) | 4204 (10.6) | 204 (5.2) | 3995 (94.8) |
| Current (pack-years Tertile 1) | 1876 (5.0) | 12 (0.6) | 1864 (99.4) |
| Current (pack-years Tertile 2) | 1759 (5.1) | 15 (1.0) | 1742 (99.0) |
| Current (pack-years Tertile 3) | 4628 (12.4) | 62 (1.6) | 4562 (98.4) |
| Missing | 34 (0.1) | 0 (0.0) | 34 (100.0) |
| Physical activity | |||
| No | 14,652 (32.1) | 342 (2.7) | 14,302 (97.3) |
| Moderate | 11,276 (30.8) | 472 (4.5) | 10,799 (95.5) |
| Vigorous | 12,083 (37.1) | 367 (3.3) | 11,705 (96.7) |
| Missing | 10 (0.0) | 0 (0.0) | 10 (100.0) |
| Medication use | |||
| Aspirin use (1999–2004) | |||
| No use | 10,718 (86.8) | 245 (2.5) | 10,457 (97.5) |
| Low use | 904 (6.5) | 67 (8.8) | 835 (91.2) |
| High use | 780 (5.4) | 54 (7.4) | 724 (92.6) |
| Missing | 172 (1.3) | 12 (7.6) | 160 (92.4) |
| Non-aspirin NSAID use (1999–2004) | |||
| No use | 11,708 (91.8) | 320 (2.9) | 11,368 (97.1) |
| Low use | 604 (5.9) | 40 (6.1) | 564 (93.9) |
| High use | 173 (1.6) | 11 (9.8) | 162 (90.2) |
| Missing | 89 (0.7) | 7 (6.9) | 82 (93.1) |
| Health status and other risk factors | |||
| Self-reported health status | |||
| Excellent/very good | 15,970 (50.1) | 632 (4.2) | 15,329 (95.8) |
| Good/fair | 20,365 (46.8) | 523 (2.9) | 19,830 (97.1) |
| Poor | 1665 (3.1) | 25 (1.3) | 1637 (98.7) |
| Missing | 21 (0.0) | 1 (5.4) | 20 (96.5) |
| Morbidity scoreb | |||
| None | 25,461 (70.6) | 727 (3.3) | 24,724 (96.7) |
| Low | 5688 (15.4) | 250 (4.5) | 5430 (95.5) |
| High | 6353 (12.9) | 184 (3.3) | 6163 (96.7) |
| Missing | 519 (1.0) | 20 (4.4) | 499 (96.5) |
| History of arthritis | |||
| No | 8664 (21.1) | 606 (7.7) | 8047 (92.3) |
| Yes | 29,276 (78.8) | 573 (2.3) | 28,690 (97.7) |
| Missing | 81 (0.2) | 2 (2.6) | 79 (97.4) |
Regular use defined by use in the month before baseline and reported usual frequency of use 20+ days/month.
Morbidity score is the linear predictor for each individual based on a Cox proportional hazards model modeling time to death as a function of comorbidities, including diabetes, kidney failure, asthma, congestive heart failure, myocardial infarction, stroke, emphysema, chronic bronchitis, liver condition, cancer, and osteoporosis.
GED, general educational development; NSAID, nonsteroidal anti-inflammatory drug.
As effect estimates were comparable in the two models and their inclusion reduced the sample size (due to the limited cycle availability), these variables were excluded from the final models. All data on the above covariates were collected at baseline.
Given prior observation of a stronger association for colorectal cancer in the early years of follow-up,17,20,26 separate secondary analyses were conducted in early and late follow-up (defined as before and after median follow-up, respectively). Specifically, in models of early follow-up, all results are censored at 8.1 years; in models of late follow-up, follow-up starts at 8.1 years, with anyone censored in the first half of follow-up not contributing to this analysis. To further disaggregate this from calendar time (and given evidence that the association for CRP may be stronger in the earlier years21,22,24,26), analyses stratified on baseline year of exposure assessment and cohort entry (1999–2006 vs. 2007–2014) were also conducted.
Secondary analyses were conducted to examine associations with cause-specific mortality, including cancer-specific, cardiovascular-specific, and other mortality. In these analyses, those who died from causes other than the cause under study were censored at date of death. Lastly, to facilitate comparisons with prior cohorts, a sensitivity analysis was conducted restricted to persons ages 50+ years at the time of baseline survey. All secondary/sensitivity analyses were conducted for the two primary exposures: regular glucosamine use and regular chondroitin use.
All analyses were weighted to account for unequal sampling probability and nonresponse and were conducted using Stata version 15.1.43
Results
Of the 38,021 adults included in this study, 1181 (3.5%) reported regular glucosamine use (Table 1) and 2.4% reported regular chondroitin use. Glucosamine use was most prevalent in the age categories of 60–69 and 70–79, with a prevalence of 8.2% in both, compared with 0.5% and 0.9% in the 20–29 and 30–39 age categories, respectively. Glucosamine use was also correlated with income: among those with a poverty-to-income ratio of greater than or equal to 4 (highest income), 5.3% of people were regular glucosamine users compared with those with a poverty-to-income ratio of <1 (lowest income), where only 1.1% of people were regular glucosamine users. Regular glucosamine use is more prevalent in non-Hispanic white, married, and higher educated adults. Glucosamine users were also more likely than nonusers to have some or high use of aspirin and other NSAIDs.
In this study, 4908 persons died over a median of 8.1 years of follow-up (max: 16.75 years), including 1086 deaths due to cancer, 840 deaths due to cardiovascular disease, and 2969 deaths due to other causes. Of these deaths, 12.2% were among those who regularly used any glucosamine, and 8.8% in those who did not (Table 2).
Table 2.
Regular Use of Glucosamine, Chondroitin, and Methylsulfonylmethane Supplements in Relation to Mortality
| Exposure | Cohort, N (weighted %) | Deaths, N (weighted %) | Age and sex-adjusted |
Multivariable-adjusteda |
||
|---|---|---|---|---|---|---|
| HR | 95% CI | HR | 95% CI | |||
| Primary exposures | ||||||
| Regular use of any glucosamineb | ||||||
| No regular use | 36,816 (96.5) | 4696 (8.8) | 1.00 | Ref. | 1.00 | Ref. |
| Regular use | 1181 (3.5) | 204 (12.2) | 0.70 | 0.59–0.83 | 1.02 | 0.86–1.21 |
| Regular use of any chondroitinb | ||||||
| No regular use | 37,178 (97.6) | 4748 (8.8) | 1.00 | Ref. | 1.00 | Ref. |
| Regular use | 826 (2.4) | 156 (13.5) | 0.74 | 0.61–0.89 | 1.04 | 0.87–1.25 |
| Secondary exposures | ||||||
| Regular use of glucosamine+chondroitinb | ||||||
| No regular use of either | 36,802 (96.5) | 4693 (8.8) | 1.00 | Ref. | 1.00 | Ref. |
| Regular use of glucosamine only | 369 (1.1) | 51 (9.4) | 0.63 | 0.46–0.85 | 0.95 | 0.69–1.31 |
| Regular use of both | 812 (2.3) | 153 (13.5) | 0.73 | 0.60–0.88 | 1.05 | 0.87–1.26 |
| Regular use of any MSMb | ||||||
| No use | 37,559 (98.7) | 4834 (8.9) | 1.00 | Ref. | 1.00 | Ref. |
| Use | 457 (1.3) | 71 (10.6) | 0.77 | 0.59–0.99 | 0.99 | 0.74–1.33 |
Adjusted for age, sex, race/ethnicity, educational status, marital status, poverty-to-income ratio, health insurance, body mass index, alcohol use, smoking, physical activity, aspirin use, nonaspirin NSAID use, self-reported heath status, morbidity score, history of arthritis, and survey cycle.
Regular use defined by use in the month before baseline and reported usual frequency of use 20+ days/month.
95% CI, 95% confidence interval; HR, hazard ratio; MSM, methylsulfonylmethane; NSAID, nonsteroidal anti-inflammatory drug.
Despite glucosamine use seemingly inversely associated with mortality in minimally adjusted models (HR: 0.70; 95% CI: 0.59–0.84), following multivariable adjustment, no association was observed (HR: 1.02; 95% CI: 0.85–1.23) (Table 2). Similar results were observed for chondroitin and joint use of glucosamine and chondroitin, with no association observed after multivariable adjustment (chondroitin: HR: 1.05; 95% CI: 0.87–1.27; glucosamine chondroitin: HR: 1.05; 95% CI: 0.86–1.28; or MSM: HR: 1.02; 95% CI: 0.74–1.40).
For cause-specific mortality, a significant inverse association was found after regular use of glucosamine/chondroitin for both cardiovascular mortality and other cause mortality rate in the minimally adjusted model (Table 3). However, after multivariable adjustment, the association did not remain statistically significant for the cardiovascular mortality (glucosamine: HR 0.72; 95% CI: 0.45–1.14, chondroitin: HR: 0.75; 95% CI: 0.46–1.21) and was null for other cause mortality rate (glucosamine: HR: 1.00; 95% CI: 0.79–1.25, chondroitin: HR: 0.98; 95% CI: 0.78–1.23). There was no significant association between use of either glucosamine or chondroitin and cancer mortality (Table 3).
Table 3.
Association Between Glucosamine and Chondroitin Use and Mortality
| Cohort, N(%) | Case, N (%) | Age and sex-adjusted, HR (95% CI) | Multivariable-adjusted,a HR (95% CI) | |
|---|---|---|---|---|
| Overall mortality | ||||
| Regular use of any glucosamineb | ||||
| No regular use | 36,816 (96.5) | 4696 (8.8) | 1.00 (Ref.) | 1.00 (Ref.) |
| Regular use | 1181 (3.5) | 204 (12.2) | 0.70 (0.59–0.83) | 1.02 (0.86–1.21) |
| Regular use of any chondroitinb | ||||
| No regular use | 37,178 (97.6) | 4748 (8.8) | 1.00 (Ref.) | 1.00 (Ref.) |
| Regular use | 826 (2.4) | 156 (13.5) | 0.74 (0.61–0.89) | 1.04 (0.87–1.25) |
| Cancer mortality | ||||
| Regular use of any glucosamineb | ||||
| No regular use | 36,816 (96.5) | 1028 (2.0) | 1.00 (Ref.) | 1.00 (Ref.) |
| Regular use | 1181 (3.5) | 55 (3.6) | 0.90 (0.64–1.25) | 1.19 (0.83–1.71) |
| Regular use of any chondroitinb | ||||
| No regular use | 37,178 (97.6) | 1040 (2.0) | 1.00 (Ref.) | 1.00 (Ref.) |
| Regular use | 826 (2.4) | 45 (4.2) | 1.03 (0.74–1.45) | 1.33 (0.92–1.90) |
| Cardiovascular mortality | ||||
| Regular use of any glucosamineb | ||||
| No regular use | 36,816 (96.5) | 815 (1.4) | 1.00 (Ref.) | 1.00 (Ref.) |
| Regular use | 1181 (3.5) | 24 (1.4) | 0.47 (0.30–0.74) | 0.72 (0.46–1.15) |
| Regular use of any chondroitinb | ||||
| No regular use | 37,178 (97.6) | 819 (1.4) | 1.00 (Ref.) | 1.00 (Ref.) |
| Regular use | 826 (2.4) | 20 (1.6) | 0.51 (0.31–0.81) | 0.76 (0.47–1.21) |
| Other mortality | ||||
| Regular use of any glucosamineb | ||||
| No regular use | 36,816 (96.5) | 2840 (5.4) | 1.00 (Ref.) | 1.00 (Ref.) |
| Regular use | 1181 (3.5) | 125 (7.2) | 0.69 (0.56–0.85) | 1.03 (0.83–1.28) |
| Regular use of any chondroitinb | ||||
| No regular use | 37,178 (97.6) | 2876 (5.4) | 1.00 (Ref.) | 1.00 (Ref.) |
| Regular use | 826 (2.4) | 91 (7.7) | 0.70 (0.56–0.87) | 1.01 (0.81–1.26) |
Adjusted for age, sex, race/ethnicity, educational status, marital status, poverty-to-income ratio, health insurance, body mass index, alcohol use, smoking, physical activity, aspirin use, nonaspirin NSAID use, self-reported heath status, morbidity score, history of arthritis, and survey cycle.
Regular use defined by use in the month before baseline and reported usual frequency of use 20+ days/month.
95% CI, 95% confidence interval; HR, hazard ratio; NSAID, nonsteroidal anti-inflammatory drug.
In follow-up time stratified models, there appeared to be some difference in early and late years. Specifically, an HR of 1.08 (95% CI: 0.90–1.31) was observed in the early years of follow-up for glucosamine, compared with 0.86 (95% CI: 0.57–1.29) in the later years of follow-up (Supplementary Table S3). Similarly, an HR of 1.15 (95% CI: 0.93–1.42) was observed for chondroitin in early follow-up, compared with 0.79 (95% CI: 0.53–1.19) in later follow-up (Supplementary Table S4). For both glucosamine and chondroitin, there was no evidence of interaction by year of study entry (p-interaction: 0.98 and p-interaction: 0.99, respectively) (Supplementary Tables S1 and S2).
In keeping with other studies, analyses were conducted restricted to those aged 50+. Results were comparable with those observed in overall models (Supplementary Table S5).
Discussion
In this nationally representative study, no significant association with all-cause mortality after multivariable adjustment was identified. Despite evidence of an association after adjustment for age and sex, the association attenuated to the null with incorporation of factors pertaining to lifestyle, demographics, and overall health measures. A similar pattern of association was observed for cause-specific mortality (mortality due to cancer, cardiovascular disease, and other causes). Further work in a larger study will be needed to better understand a potentially suggestive inverse (nonsignificant) association with cardiovascular mortality.
Our observation of no association between glucosamine and mortality contrasted with prior studies, where a significant inverse relationship was found.19,20,22,24 It is likely that the difference in comparison with the prior NHANES study22 reflects more complete adjustment for confounding factors, as this study included a less extensive panel of covariates (without inclusion of variables such as BMI, self-reported health status, or income); however, it seems unlikely that this explains the difference with the prior VITAL studies,19,24 which also included a comprehensive panel of covariates. Notably, the population prevalence of use is much lower in this cohort (3.5%) versus 20.9% in VITAL. It is possible that this reflects differences in how use was captured and what it represents.
A difference in prevalence may also account for the difference between this study and the U.K. Biobank study, where 19.1% of the study population used glucosamine, which is regulated as a prescription drug in the United Kingdom.20 Initial consideration suggested that this difference was due to age differences in the population (20+ here, vs. ages 50+ in VITAL and 40+ in the U.K. Biobank study), with younger adults having a lower risk of death due to cancer and cardiovascular disease and thus a lower potential immediate impact of an anti-inflammatory supplement. However, in post hoc sensitivity analyses, the overall effect estimates for glucosamine and chondroitin and the overall mortality were comparable in analyses of adults ages 50+ years.
In this study, no significant association between glucosamine/chondroitin and specific causes of mortality was found, in contrast to previous studies. In prior work, Bell et al. found a significant risk reduction for both death from cancer and respiratory disease.19 This study was unable to evaluate respiratory-specific mortality rate, although no significant association with cancer mortality was observed, in contrast to the study by Bell et al. It is unlikely to be explained by difference in covariate adjustment, as both studies adjusted for a similar panel of covariates. One difference is that in the study by Bell et al., participants were aged 50–76 at baseline, in comparison with this study where adults older than 20 years were included.
Given the limitations in power in the present study, it was not possible to examine cause-specific mortality alongside age restriction, to further evaluate this. It should also be noted that in other studies, glucosamine/chondroitin use has been observed to be associated with reduced risk of colorectal cancer and lung cancer, and thus, the finding of no association with cancer mortality in this study is surprising.17,18,20,23,25,26 Several other studies examined the association between glucosamine use and cardiovascular disease and found a significant inverse relationship,21,22 although both the King and Xiang and Ma et al. studies adjusted for far fewer potential confounders. A nonsignificant, inverse association was observed in this study. Further work in a well-powered study will be needed to better understand this association.
Results stratified by follow-up time suggested a nonsignificant association between glucosamine/chondroitin and decreased mortality in the later years of follow-up, with no association observed in early-follow-up. No difference in association was observed by calendar time of cohort entry. This suggests that these differences reflect the increased length of follow-up time, rather than any change in the exposure definition. Previous studies have noted an association between glucosamine and chondroitin and colorectal cancer, although they found decreased inflammatory biomarkers in the earlier years of follow-up.26,41 There was no clear reason identified for this association.
There were several limitations in this study. NHANES excludes those residing in institutional settings. Given the data showing increased glucosamine use in the elderly and that as of 2017, there are 1.5 million senior citizens in America residing in institutionalized settings, this is an important limitation.44 Although 3.5% of the population were regular users of glucosamine, only 2.4% used chondroitin regularly. Given the small numbers of those using glucosamine alone and chondroitin alone, it was not possible to compare glucosamine/chondroitin to glucosamine alone and chondroitin alone in the subsequent secondary/sensitivity analyses due to limited power.
In this study, regular use was defined by those who reported use of glucosamine on 20 or more days in the 30 days before survey administration, and thus, exposure is assessed by prevalent use, which could theoretically result in selection bias, although this is less likely in the context of a cancer outcome.45 It is possible that these users may not have been consistent or long-term users through the etiologically relevant time-period, resulting in measurement error and possibly attenuating results toward the null. Similarly, it is possible that covariate status may have changed throughout follow-up, leading to measurement error in covariates and residual confounding. In the United States, where glucosamine is available over the counter, the confounding variables might vary in comparison with a population where glucosamine is used by prescription (as is the case in many European countries); that said, most prior work has been similarly conducted in U.S. populations.
There are several strengths in this study. It uniquely explores the impact of additional covariates (including demographics, physical health, medication use, a morbidity score, and self-reported health status) on the association between regular glucosamine use and mortality and found no significant association, suggesting that the association identified in existing studies may be secondary to residual unmeasured confounders. Extensive steps were taken to address confounding regarding healthy user bias, including adjusting for self-reported health status. The study also leveraged a nationally representative survey with a large sample size and subsequently highly powered analysis. Finally, the study was able to examine stratifications by study time and follow-up time to address questions about potential heterogeneity in association with time.
In this nationally representative study of U.S. adults, no significant associations between glucosamine/chondroitin and all-cause and cause-specific mortality were identified in contrast to prior studies, including a prior NHANES study. However, a suggestive, nonsignificant inverse association with cardiovascular mortality was seen; given the limited power to explore cause-specific mortality, future well-powered studies will be needed to better understand the potential association with cardiovascular-specific mortality.
Supplementary Material
Authors' Contributions
J.B.: writing—original draft and writing—reviewing and editing; K.O.: methodology, data curation, formal analysis, and writing—reviewing and editing; D.K.: investigation, methodology, data curation, formal analysis, and writing—reviewing and editing; M.D.: writing—reviewing and editing; S.L.N.: investigation and writing—reviewing and editing; E.D.K.: conceptualization, supervision, project administration, data curation, methodology, funding acquisition, and writing—reviewing and editing.
Author Disclosure Statement
The authors have no conflict of interest to declare.
Funding Information
This work was supported by grant P30CA008748 from the National Cancer Institute (NCI) as well as the Geoffrey Beene Cancer Research Center at Memorial Sloan Kettering Cancer Center.
Supplementary Material
References
- 1. Qato DM, Alexander GC, Conti RM, et al. Use of prescription and over-the-counter medications and dietary supplements among older adults in the United States. JAMA 2008;300(24):2867–2878; doi: 10.1001/jama.2008.892 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Kantor ED, Rehm CD, Du M, et al. Trends in dietary supplement use among US adults from 1999–2012. JAMA 2016;316(14):1464–1474; doi: 10.1001/jama.2016.14403 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Anderson JW, Nicolosi RJ, Borzelleca JF. Glucosamine effects in humans: A review of effects on glucose metabolism, side effects, safety considerations and efficacy. Food Chem Toxicol 2005;43(2):187–201; doi: 10.1016/j.fct.2004.11.006. [DOI] [PubMed] [Google Scholar]
- 4. Rindone JP, Hiller D, Collacott E, et al. Randomized, controlled trial of glucosamine for treating osteoarthritis of the knee. West J Med 2000;172(2):91–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Reginster J-Y, Veronese N. Highly purified chondroitin sulfate: A literature review on clinical efficacy and pharmacoeconomic aspects in osteoarthritis treatment. Aging Clin Exp Res 2021;33(1):37–47; doi: 10.1007/s40520-020-01643-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Graham DJ, Campen D, Hui R, et al. Risk of acute myocardial infarction and sudden cardiac death in patients treated with cyclo-oxygenase 2 selective and non-selective non-steroidal anti-inflammatory drugs: Nested case-control study. Lancet 2005;365(9458):475–481; doi: 10.1016/S0140-6736(05)17864-7 [DOI] [PubMed] [Google Scholar]
- 7. Bombardier C, Laine L, Reicin A, et al. Comparison of upper gastrointestinal toxicity of rofecoxib and naproxen in patients with rheumatoid arthritis. N Engl J Med 2000;343(21):1520–1528; doi: 10.1056/NEJM200011233432103 [DOI] [PubMed] [Google Scholar]
- 8. Jordan KM, Arden NK, Doherty M, et al. EULAR Recommendations 2003: An evidence based approach to the management of knee osteoarthritis: Report of a Task Force of the Standing Committee for International Clinical Studies Including Therapeutic Trials (ESCISIT). Ann Rheum Dis 2003;62(12):1145–1155; doi: 10.1136/ard.2003.011742 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Kerr SJ, Rowett DS, Sayer GP, et al. All-cause mortality of elderly Australian veterans using COX-2 selective or non-selective NSAIDs: A longitudinal study. Br J Clin Pharmacol 2011;71(6):936–942; doi: 10.1111/j.1365-2125.2010.03702.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Gislason GH, Rasmussen JN, Abildstrom SZ, et al. Increased mortality and cardiovascular morbidity associated with use of nonsteroidal anti-inflammatory drugs in chronic heart failure. Arch Intern Med 2009;169(2):141–149; doi: 10.1001/archinternmed.2008.525 [DOI] [PubMed] [Google Scholar]
- 11. Olsen A-MS, Fosbøl EL, Lindhardsen J, et al. Long-term cardiovascular risk of nonsteroidal anti-inflammatory drug use according to time passed after first-time myocardial infarction: A nationwide cohort study. Circulation 2012;126(16):1955–1963; doi: 10.1161/CIRCULATIONAHA.112.112607 [DOI] [PubMed] [Google Scholar]
- 12. McAlindon TE, LaValley MP, Gulin JP, et al. Glucosamine and chondroitin for treatment of osteoarthritis: A systematic quality assessment and meta-analysis. JAMA 2000;283(11):1469–1475; doi: 10.1001/jama.283.11.1469 [DOI] [PubMed] [Google Scholar]
- 13. Glyn-Jones S, Palmer AJR, Agricola R, et al. Osteoarthritis. Lancet 2015;386(9991):376–387; doi: 10.1016/S0140-6736(14)60802-3 [DOI] [PubMed] [Google Scholar]
- 14. Herrero-Beaumont G, Ivorra JAR, del Carmen Trabado M, et al. Glucosamine sulfate in the treatment of knee osteoarthritis symptoms: A randomized, double-blind, placebo-controlled study using acetaminophen as a side comparator. Arthritis Rheum 2007;56(2):555–567; doi: 10.1002/art.22371 [DOI] [PubMed] [Google Scholar]
- 15. Wandel S, Jüni P, Tendal B, et al. Effects of glucosamine, chondroitin, or placebo in patients with osteoarthritis of hip or knee: Network meta-analysis. BMJ 2010;341:c4675; doi: 10.1136/bmj.c4675 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Clegg DO, Reda DJ, Harris CL, et al. Glucosamine, chondroitin sulfate, and the two in combination for painful knee osteoarthritis. N Engl J Med 2006;354(8):795–808; doi: 10.1056/NEJMoa052771 [DOI] [PubMed] [Google Scholar]
- 17. Kantor E, Lampe J, Peters U, et al. Use of glucosamine and chondroitin supplements and risk of colorectal cancer. Cancer Causes Control 2013;24(6):1137–1146; doi: 10.1007/s10552-013-0192-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Satia JA, Littman A, Slatore CG, et al. Associations of herbal and specialty supplements with lung and colorectal cancer risk in the VITamins And Lifestyle (VITAL) Study. Cancer Epidemiol Biomarkers Prev 2009;18(5):1419–1428; doi: 10.1158/1055-9965.EPI-09-0038 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Bell GA, Kantor ED, Lampe JW, et al. Use of glucosamine and chondroitin in relation to mortality. Eur J Epidemiol 2012;27(8):593–603; doi: 10.1007/s10654-012-9714-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Li Z-H, Gao X, Chung VC, et al. Associations of regular glucosamine use with all-cause and cause-specific mortality: A large prospective cohort study. Ann Rheum Dis 2020;79(6):829–836; doi: 10.1136/annrheumdis-2020-217176 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Ma H, Li X, Sun D, et al. Association of habitual glucosamine use with risk of cardiovascular disease: Prospective study in UK Biobank. BMJ 2019;365:l1628; doi: 10.1136/bmj.l1628 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. King DE, Xiang J. Glucosamine/chondroitin and mortality in a US NHANES cohort. J Am Board Fam Med 2020;33(6):842–847; doi: 10.3122/jabfm.2020.06.200110 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Brasky TM, Lampe JW, Slatore CG, et al. Use of glucosamine and chondroitin and lung cancer risk in the VITamins And Lifestyle (VITAL) cohort. Cancer Causes Control 2011;22(9):1333–1342; doi: 10.1007/s10552-011-9806-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Pocobelli G, Kristal AR, Patterson RE, et al. Total mortality risk in relation to use of less-common dietary supplements. Am J Clin Nutr 2010;91(6):1791–1800; doi: 10.3945/ajcn.2009.28639 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Kantor ED, Newton CC, Giovannucci EL, et al. Glucosamine use and risk of colorectal cancer: Results from the Cancer Prevention Study II Nutrition Cohort. Cancer Causes Control 2018;29(3):389–397; doi: 10.1007/s10552-018-1003-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Kantor ED, O'Connell K, Liang PS, et al. Glucosamine use and risk of colorectal cancer: Results from UK Biobank. Cancer Epidemiol Biomarkers Prev 2022;31(3):647–653; doi: 10.1158/1055-9965.EPI-21-1171 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Kantor ED, Zhang X, Wu K, et al. Use of glucosamine and chondroitin supplements in relation to risk of colorectal cancer: Results from the Nurses' Health Study and Health Professionals follow-up study. Int J Cancer 2016;139(9):1949–1957; doi: 10.1002/ijc.30250 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Li G, Zhang X, Liu Y, et al. Relationship between glucosamine use and the risk of lung cancer: Data from a nationwide prospective cohort study. Eur Respir J 2022;59(3):2101399; doi: 10.1183/13993003.01399-2021 [DOI] [PubMed] [Google Scholar]
- 29. Largo R, Alvarez-Soria MA, Díez-Ortego I, et al. Glucosamine inhibits IL-1β-induced NFκB activation in human osteoarthritic chondrocytes. Osteoarthritis Cartilage 2003;11(4):290–298; doi: 10.1016/S1063-4584(03)00028-1 [DOI] [PubMed] [Google Scholar]
- 30. Wu Y-L, Kou YR, Ou H-L, et al. Glucosamine regulation of LPS-mediated inflammation in human bronchial epithelial cells. Eur J Pharmacol 2010;635(1):219–226; doi: 10.1016/j.ejphar.2010.02.044 [DOI] [PubMed] [Google Scholar]
- 31. Yomogida S, Hua J, Sakamoto K, et al. Glucosamine suppresses interleukin-8 production and ICAM-1 expression by TNF-α-stimulated human colonic epithelial HT-29 cells. Int J Mol Med 2008;22(2):205–211; doi: 10.3892/ijmm_00000010 [DOI] [PubMed] [Google Scholar]
- 32. Gouze J-N, Bordji K, Gulberti S, et al. Interleukin-1β down-regulates the expression of glucuronosyltransferase I, a key enzyme priming glycosaminoglycan biosynthesis: Influence of glucosamine on interleukin-1β–mediated effects in rat chondrocytes. Arthritis Rheum 2001;44(2):351–360; doi: [DOI] [PubMed] [Google Scholar]
- 33. Hua J, Sakamoto K, Kikukawa T, et al. Evaluation of the suppressive actions of glucosamine on the interleukin-1β-mediated activation of synoviocytes. Inflamm Res 2007;56(10):432–438; doi: 10.1007/s00011-007-7020-7 [DOI] [PubMed] [Google Scholar]
- 34. Rajapakse N, Kim M-M, Mendis E, et al. Inhibition of inducible nitric oxide synthase and cyclooxygenase-2 in lipopolysaccharide-stimulated RAW264.7 cells by carboxybutyrylated glucosamine takes place via down-regulation of mitogen-activated protein kinase-mediated nuclear factor-κB signaling. Immunology 2008;123(3):348–357; doi: 10.1111/j.1365-2567.2007.02683.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Kantor ED, Lampe JW, Navarro SL, et al. Associations between glucosamine and chondroitin supplement use and biomarkers of systemic inflammation. J Altern Complement Med 2014;20(6):479–485; doi: 10.1089/acm.2013.0323 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Kantor ED, Lampe JW, Vaughan TL, et al. Association between use of specialty dietary supplements and C-reactive protein concentrations. Am J Epidemiol 2012;176(11):1002–1013; doi: 10.1093/aje/kws186 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Navarro SL, White E, Kantor ED, et al. Randomized trial of glucosamine and chondroitin supplementation on inflammation and oxidative stress biomarkers and plasma proteomics profiles in healthy humans. PLoS One 2015;10(2):e0117534; doi: 10.1371/journal.pone.0117534 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Nakamura H, Nishioka K. Effects of glucosamine/chondroitin supplement on osteoarthritis: Involvement of PGE2 and YKL-40. J Rheum Jt Surg 2002;21(2):175–184; doi: 10.11551/jsjd1982.21.175 [DOI] [Google Scholar]
- 39. Marsik C, Kazemi-Shirazi L, Schickbauer T, et al. C-reactive protein and all-cause mortality in a large hospital-based cohort. Clin Chem 2008;54(2):343–349; doi: 10.1373/clinchem.2007.091959 [DOI] [PubMed] [Google Scholar]
- 40. Sullivan DH, Roberson PK, Johnson LE, et al. Association between inflammation-associated cytokines, serum albumins, and mortality in the elderly. J Am Med Dir Assoc 2007;8(7):458–463; doi: 10.1016/j.jamda.2007.04.004 [DOI] [PubMed] [Google Scholar]
- 41. Kantor ED, O'Connell K, Du M, et al. Glucosamine and chondroitin use in relation to C-reactive protein concentration: Results by supplement form, formulation, and dose. J Altern Complement Med 2021;27(2):150–159; doi: 10.1089/acm.2020.0283 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Anonymous. NCHS Data Linkage—Mortality Data. 2022. Available from: https://www.cdc.gov/nchs/data-linkage/mortality.htm [Last accessed: March 4, 2022].
- 43. National Center for Health Statistics (U.S.), (ed). National Health and Nutrition Examination Survey: Analytic Guidelines, 1999–2010. Vital and Health Statistics. Series 2 Number 161. U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics: Hyattsville, MD; 2013. [Google Scholar]
- 44. Anonymous. Profile of Older Americans | ACL Administration for Community Living. n.d. Available from: https://acl.gov/aging-and-disability-in-america/data-and-research/profile-older-americans [Last accessed: July 20, 2022].
- 45. Suissa K, Hudson M, Suissa S. Glucosamine and lower mortality and cancer incidence: Selection bias in the observational studies. Pharmacoepidemiol Drug Saf 2022;31(12):1272–1279; doi: 10.1002/pds.5535 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.