DTI-Derived Evaluation of Glymphatic System Function in Veterans with Chronic Multisymptom Illness

Glymphatic System and Neurotoxic Exposures

The commonly known condition of GWI was described by the Centers for Disease Control and Prevention¹³ as an unexplained CMI that persistently presents in approximately 30%–44% of the nearly 700,000 US military personnel who served in the 1990–1991 Persian Gulf War. Postdeployment CMI appears to be more common in modern wars than previously appreciated, with some studies reporting CMI²² and brainstem alterations²⁰ in veterans who served in Afghanistan and/or Iraq Wars. Veterans who deployed to Gulf War and Afghanistan/Iraq Wars had similar environmental exposures, including sand dust, oil well fires, burn pits, chemical weapons (eg, sarin gas), chemical nerve agent antidotes (pyridostigmine bromide or nerve agent pyridostigmine pretreatment), pesticide uses, vaccines not previously used by the military, depleted uranium munitions, as well as hearing chemical alarms or wearing mission-oriented protective posture gear.²³ GWI is reported to be strongly associated with using pyridostigmine bromide pills and wearing mission-oriented protective posture gear.²⁴ Furthermore, veterans’ exposures to Scud missiles, and to chemical and biologic warfare agents were associated with major CMI symptoms.²⁵ Although exposures to these toxic substances have been considered as a possible cause for acute symptoms during the wars²⁶ and an initial contributing factor for CMI, they are unlikely to explain the persistent chronic conditions that occur, usually months after returning from deployment. Studies in animal models of GWI suggest that war-related stress, head trauma, and repeated pyridostigmine bromide administration are associated with disruption of the BBB.²⁷ Thus, soldiers with impaired glymphatic clearance may have experienced long-term, aberrant accumulation of neurotoxic substances, potentially leading to multiple chronic symptoms, but further research is required to establish this link.

DTI-ALPS and Aging

Animal and human studies have shown that glymphatic activity decreases in elderly populations.²⁸ The reduced clearance of metabolic waste such as amyloid-β and τ aggregates may contribute to aging and cognitive impairment.⁵ In this study, the observation of a strong association between age and ALPS in the HC group is consistent with the notion that aging adversely affects glymphatic function. On the other hand, sleep problems often worsen with normal aging, as well as in CMI and other conditions commonly seen in postdeployment veterans such as PTSD, leading to disrupted non-REM sleep stage and a decline of glymphatic clearance.²⁹ In this study, we observed a weaker effect of age-related glymphatic dysfunction among the CMI group, in comparison with the strong effect in HCs. This observation is not surprising, because most of the veterans were in young- and mid-adulthood where aging has less impact on brain health. Therefore, the glymphatic impairment seen at this age range is more likely due to CMI pathology than aging.

We observed a higher ALPS in female HCs than in male HCs. This finding is in line with previously reported sex differences in glymphatic function measured by ALPS.³⁰ Although a previous study suggested that this sex effect may be confounded by head size,³¹ our data clearly showed that sex differences in ALPS remained significant after adjusting for total intracranial volume. As the number of women in the armed forces continues to rise, studying sex differences in glymphatic function may offer important insights for subsequent women veterans’ health.

DTI-ALPS and Sleep

Most glymphatic clearance occurs during sleep. During sleep, the efficiency of amyloid-beta clearance has been found to double, and conversely, sleep deprivation leads to a reduction in the clearance of metabolic waste.⁵ Using DTI-ALPS, studies have implicated glymphatic dysfunction in idiopathic REM sleep behavior disorder,³² obstructive sleep apnea (OSA),³³ and non-REM sleep,¹⁰ suggesting that poor sleep quality is highly relevant to glymphatic dysfunctions. Sleep disturbances are very common in veterans from modern wars and postcombat CMI. Our data showed that about 96% of veterans with CMI had insomnia. Although a significant correlation between ALPS and PSQI-glob was not found in this study, we observed that ALPS was substantially lower in veterans with insomnia compared with veterans with minor sleep problems (Fig 4). Furthermore, we also observed that sleep disturbance, one of the PSQI components (particularly items related to snoring and breathing difficulties), correlated with glymphatic dysfunction. These results are consistent with a published study¹⁰ of elderly civilians that found ALPS was associated with non-REM sleep and apnea-hypopnea but not PSQI. Other evidence includes sleep-disordered breathing-related glymphatic circulation issues³⁴ and decreased glymphatic clearance by OSA.³⁵ There is a broad consensus that sleep plays an important role in maintaining normal glymphatic system function, and sleep disturbances conversely result in impaired glymphatic functions and accumulation of neural wastes and toxins.³⁶

DTI-ALPS and Pain

The glymphatic system is known to be regulated by sleep and norepinephrine. Glymphatic clearance is optimal under conditions of good sleep and blockade of adrenergic signaling.³⁷ Conversely, norepinephrine, an arousal-mediation neurotransmitter, promotes wakefulness and reduces glymphatic system function by decreasing the size of the interstitial space.³⁸ Other studies on persistent pain conditions and headaches suggest that impaired glymphatic drainage can lead to excessive accumulation of calcitonin gene-related peptide (CGRP), a mediator that aggravates headaches and chronic migraines.³⁹ Our findings of a correspondence between higher chronic pain intensity and poorer glymphatic clearance are consistent with these reports. Furthermore, animal studies also suggest that impaired glymphatic function is a potential driver of pain chronification.⁴⁰ In chronic pain, the norepinephrine levels in the brain increase due to elevated locus coeruleus–derived noradrenergic tone,³ which could result in inactivated CSF drainage.

DTI-ALPS and Fatigue

Less is known about whether the glymphatic function is impaired in chronic fatigue. A recent study has proposed that the “post-COVID-19 fatigue syndrome” may be caused by congestion of the glymphatic system with subsequent toxic buildup within the central nervous system.⁴¹ Interestingly, similarities in symptomatology between COVID-19 and GWI have been noted by several authors.⁴² In our study, we observed a strong relationship between the degree of chronic fatigue and impaired glymphatic system function. Postcombat physical exhaustion could in part stem from glymphatic dysfunction due to sleep disturbances, disrupted neurologic restoration and performance induced by the accumulation of toxic and metabolic wastes.

DTI-ALPS and CMI

The CFS score represents a wide spectrum of symptoms, including, but not limited to, fatigue, headaches, joint and muscle pain sleep problems, memory and mood issues, sensitivity in lymph node, respiratory, gastrointestinal, and sensory problems. In this study, we used a summarized CFS severity score to indicate veterans’ CMI severity. As discussed above, glymphatic impairment may be directly related to impaired clearance of neurotoxic substances (that veterans might have been exposed to), CGRP, and metabolic wastes, and is also indirectly related to sleep disturbances that lead to decreased glymphatic function. Stress, particularly chronic stress, can affect glymphatic clearance as it reduces glymphatic activity during sleep⁶ and impairs the function of aquaporin-4.⁴³ These negative effects on the glymphatic system by stress could partially explain the glymphatic failure in CMI, because chronic PTSD is commonly comorbid with CMI. In addition, reduced glymphatic drainage of metabolic waste such as amyloid-β and τ has been considered relevant to memory and cognitive impairments,⁴⁴ and postcombat TBI.⁶ Some other concerns include glymphatic impairment as an immune dysfunction causing toxic buildup in CSF⁴⁵; low-grade inflammation within CSF might facilitate diffuse stagnation of flow in the interstitial and perivascular spaces.⁴⁶

Considering the above, it is conceivable that disturbances in the glymphatic drainage system may contribute to the pathogenesis of CMI. In this study, we did not observe significant differences in ALPS between veterans with and without comorbidities (such as TBI, PTSD, and depression) However, a significant reduction in ALPS was found in veterans with comorbid conditions compared with HCs (Fig 4). These findings suggest that glymphatic dysfunction may not be associated with a single symptom or condition but rather with systemic changes.

DTI-ALPS and Brain Volume Measures

Given that the autonomic nervous system is under the control of the brainstem, disruption or shrinkage of the brainstem, particularly the medulla and midbrain, is considered a key feature of GWI.²⁰ In this study, we found significant associations between lower DTI-ALPS and smaller volumes in the medulla, which we previously identified as a particularly vulnerable structure in GWI with respect to GWI-induced sleep and pain problems.²⁰ Further investigation by using brainstem imaging in CMI/GWI is warranted.

Limitations

There are important considerations when interpreting DTI-ALPS. Table 1 provides a comparison between DTI-ALPS and neuroimaging of CSF dynamics. Intrathecal injection of CSF tracer is still the standard for measuring glymphatic clearance⁴⁷ but is not practical due to its invasiveness. Compared with methods that examine the dynamic flow of CSF-interstitial fluid exchange, the DTI-ALPS index was proposed to measure water movement along the directions of the PVS and applied in regions of interest in periventricular regions parallel to PVS. The distinct spatial and temporal differences between DTI and CSF indicate that the DTI-ALPS index should not be viewed as a direct marker of glymphatic clearance.⁴⁸ However, DTI-ALPS has been shown to correlate with other measures of CSF dynamics,⁴⁹ as well as noninvasive methods (eg, measuring PVS burden)⁵⁰ that are proposed to reflect the status of the glymphatic system. An ongoing study of the same veteran population from our group has found that DTI-ALPS correlates with a greater PVS burden, especially with the number of enlarged cerebral PVSs. Therefore, as we described in the introduction, the merit of using DTI-ALPS in clinical application should not be ignored. There are technical problems affecting DTI-ALPS. The accuracy of DTI-ALPS may be affected not only by local tissue complexity, but also by head orientation during MRI scans. As we illustrate in the Supplemental Data, an over-tilted brain orientation or crossing-fiber tissue, can change the principal diffusivity in the fiber ROIs, leading to incorrect ALPS calculations. These challenges should be considered in future DTI-ALPS analyses. There are limitation on acquiring DTI during sleep. Some imaging studies of the CSF dynamics in anesthetized rodent brains suggest that the glymphatic system is active during sleep and inactive during wakefulness.⁴ However, more recent evidence casts doubt on the validity of this strict dichotomy⁵¹ and rather points to the more likely scenario of a gradient ranging between more and less active. Interestingly, to our knowledge there has not been any study evaluating the within-subject sleep/wake correlation of DTI-ALPS. In contrast, many studies measuring DTI-ALPS in the human brain during wakefulness have shown strong associations between glymphatic function and the accumulation of brain wastes (eg, amyloid-β). Therefore DTI-ALPS appears to be a reliable proxy for glymphatic function even during wakefulness. Furthermore, we did not test medication effects in this study. In fact, little is currently known about pharmacologic interactions with the glymphatic system other than sleep medication. Regarding the complexity of confounding factors, because of insufficient information from the healthy cohorts, we were unable to include some clinical variables (eg, alcohol use, diabetes mellitus, etc) as confounding factors. The absence of potential confounders may amplify group differences when compared with controls. Additionally, we could not include some white matter variables (eg, Fazekas scale, fractional anisotropy, and mean diffusivity) as confounders, because they exhibited collinearity, making it difficult to interpret the regression model’s results. Future research projects will be needed to reduce confounders by recruiting participants with less demographic heterogeneity and consistent clinical information and imaging protocols.