Advertisement

The bi-directional relationship between post-traumatic stress disorder and obstructive sleep apnea and/or insomnia in a large U.S. military cohort

  • Evan D. Chinoy
    Affiliations
    Sleep, Tactical Efficiency, and Endurance Laboratory, Warfighter Performance Department, Naval Health Research Center, San Diego, California, USA

    Leidos, Inc., San Diego, California, USA
    Search for articles by this author
  • Felicia R. Carey
    Affiliations
    Leidos, Inc., San Diego, California, USA

    Deployment Health Research Department, Naval Health Research Center, San Diego, California, USA
    Search for articles by this author
  • Claire A. Kolaja
    Affiliations
    Leidos, Inc., San Diego, California, USA

    Deployment Health Research Department, Naval Health Research Center, San Diego, California, USA
    Search for articles by this author
  • Isabel G. Jacobson
    Affiliations
    Leidos, Inc., San Diego, California, USA

    Deployment Health Research Department, Naval Health Research Center, San Diego, California, USA
    Search for articles by this author
  • Adam D. Cooper
    Affiliations
    Sleep, Tactical Efficiency, and Endurance Laboratory, Warfighter Performance Department, Naval Health Research Center, San Diego, California, USA

    Deployment Health Research Department, Naval Health Research Center, San Diego, California, USA

    Innovative Employee Solutions, San Diego, California, USA
    Search for articles by this author
  • Rachel R. Markwald
    Correspondence
    Corresponding author: Rachel R. Markwald, Sleep, Tactical Efficiency, and Endurance Laboratory, Warfighter Performance Department, Naval Health Research Center, 140 Sylvester Road, San Diego, CA 92106, USA.
    Affiliations
    Sleep, Tactical Efficiency, and Endurance Laboratory, Warfighter Performance Department, Naval Health Research Center, San Diego, California, USA
    Search for articles by this author
Open AccessPublished:September 24, 2022DOI:https://doi.org/10.1016/j.sleh.2022.07.005

      Abstract

      Objectives

      Determine if a bi-directional relationship exists between the development of sleep disorders (obstructive sleep apnea [OSA] and/or insomnia) and existing post-traumatic stress disorder (PTSD), and vice versa; and examine military-related factors associated with these potential relationships.

      Design

      Longitudinal analyses of a prospective representative U.S. military cohort.

      Participants

      Millennium Cohort Study responders in 2011-2013 (Time 1 [T1]) and 2014-2016 (Time 2 [T2]) without insomnia or OSA at T1 (N = 65,915) or without PTSD at T1 (N = 71,256).

      Measurements

      Provider-diagnosed OSA, self-reported items for insomnia, provider-diagnosed PTSD, and current PTSD symptoms were assessed at T1 and T2. Adjusted multivariable models identified military-related factors associated with new-onset PTSD in those with OSA and/or insomnia, and vice versa.

      Results

      Self-reported history of provider-diagnosed PTSD without current symptoms at T1 was associated with new-onset OSA only and comorbid OSA/insomnia at T2, while current PTSD symptoms and/or diagnosis was associated with new-onset insomnia only. OSA/insomnia at T1 was consistently associated with newly reported PTSD symptoms or diagnosis except that insomnia only was not associated with newly reported provider-diagnosed PTSD. Military-related risk factors significantly associated with the bi-directional relationship for new-onset PTSD or OSA/insomnia included prior deployment with higher combat exposure and recent separation from the military; being an officer was protective for both outcomes.

      Conclusions

      In this large military cohort, findings suggest that PTSD and OSA and/or insomnia are bi-directionally predictive for their development, which was sometimes revealed by health care utilization. Relevant military-related risk factors should be considered in efforts to prevent or treat PTSD and/or sleep disorders.

      Keywords

      Introduction

      Post-traumatic stress disorder (PTSD) is among the most serious conditions affecting service members and veterans, with high rates over the past 2 decades since the increase in U.S. military combat operations overseas.
      • Gates MA
      • Holowka DW
      • Vasterling JJ
      • Keane TM
      • Marx BP
      • Rosen RC.
      Posttraumatic stress disorder in veterans and military personnel: epidemiology, screening, and case recognition.
      • Walter KH
      • Levine JA
      • Highfill-McRoy RM
      • Navarro M
      • Thomsen CJ.
      Prevalence of posttraumatic stress disorder and psychological comorbidities among U.S. active duty service members, 2006-2013.
      • Judkins JL
      • Moore BA
      • Collette TL
      • Hale WJ
      • Peterson AL
      • Morissette SB.
      Incidence rates of posttraumatic stress disorder over a 17-year period in active duty military service members.
      Those with PTSD may suffer many comorbid problems after the initial trauma that worsen their physical and mental health for years afterward.
      • Walter KH
      • Levine JA
      • Highfill-McRoy RM
      • Navarro M
      • Thomsen CJ.
      Prevalence of posttraumatic stress disorder and psychological comorbidities among U.S. active duty service members, 2006-2013.
      ,
      • Ramchand R
      • Rudavsky R
      • Grant S
      • Tanielian T
      • Jaycox L.
      Prevalence of, risk factors for, and consequences of posttraumatic stress disorder and other mental health problems in military populations deployed to Iraq and Afghanistan.
      Common symptoms of PTSD include mental replay of the traumatic events, heightened stress response, and altered brain activity.
      • Germain A
      • McKeon AB
      • Campbell RL.
      Sleep in PTSD: conceptual model and novel directions in brain-based research and interventions.
      • Laxminarayan S
      • Wang C
      • Ramakrishnan S
      • et al.
      Alterations in sleep electroencephalography synchrony in combat-exposed veterans with post-traumatic stress disorder.
      • Maguire DG
      • Ruddock MW
      • Milanak ME
      • Moore T
      • Cobice D
      • Armour C.
      Sleep, a governor of morbidity in PTSD: a systematic review of biological markers in PTSD-related sleep disturbances.
      • Richards A
      • Kanady JC
      • Neylan TC.
      Sleep disturbance in PTSD and other anxiety-related disorders: an updated review of clinical features, physiological characteristics, and psychological and neurobiological mechanisms.
      These symptoms can manifest at night, disrupt bedtime routines, and negatively impact sleep, leading to a variety of sleep disorders.
      • Foster SN
      • Brock MS
      • Hansen S
      • et al.
      Sleep disorders related to deployment in active duty service members and veterans.
      ,
      • Elliott JE
      • Opel RA
      • Pleshakov D
      • et al.
      Posttraumatic stress disorder increases the odds of REM sleep behavior disorder and other parasomnias in Veterans with and without comorbid traumatic brain injury.
      With the demanding and stressful conditions required of military service (eg, continuous operations day and night, frequent travel, intensive training, long-duration work shifts, and uncomfortable sleeping conditions), regularly achieving healthy sleep while actively serving, especially during or after deployment, is challenging.
      • Foster SN
      • Brock MS
      • Hansen S
      • et al.
      Sleep disorders related to deployment in active duty service members and veterans.
      ,
      • Good CH
      • Brager AJ
      • Capaldi VF
      • Mysliwiec V.
      Sleep in the United States military.
      Thus, the conditions inherent to military service may contribute to high rates of sleep disorders in military populations.
      • Foster SN
      • Brock MS
      • Hansen S
      • et al.
      Sleep disorders related to deployment in active duty service members and veterans.
      ,
      • Good CH
      • Brager AJ
      • Capaldi VF
      • Mysliwiec V.
      Sleep in the United States military.
      • Mysliwiec V
      • Gill J
      • Lee H
      • et al.
      Sleep disorders in US military personnel: a high rate of comorbid insomnia and obstructive sleep apnea.
      • Moore BA
      • Tison LM
      • Palacios JG
      • Peterson AL
      Mysliwiec V. Incidence of insomnia and obstructive sleep apnea in active duty United States military service members.
      Beyond the essential role sleep plays in alertness and performance, sleep also impacts short- and long-term health outcomes, with insufficient or poor-quality sleep raising the risk for physical and mental health disorders, including PTSD.
      • Germain A
      • McKeon AB
      • Campbell RL.
      Sleep in PTSD: conceptual model and novel directions in brain-based research and interventions.
      ,
      • Maguire DG
      • Ruddock MW
      • Milanak ME
      • Moore T
      • Cobice D
      • Armour C.
      Sleep, a governor of morbidity in PTSD: a systematic review of biological markers in PTSD-related sleep disturbances.
      ,
      • Richards A
      • Kanady JC
      • Neylan TC.
      Sleep disturbance in PTSD and other anxiety-related disorders: an updated review of clinical features, physiological characteristics, and psychological and neurobiological mechanisms.
      ,
      • Good CH
      • Brager AJ
      • Capaldi VF
      • Mysliwiec V.
      Sleep in the United States military.
      ,
      • McLay RN
      • Klam WP
      • Volkert SL.
      Insomnia is the most commonly reported symptom and predicts other symptoms of post-traumatic stress disorder in US service members returning from military deployments.
      Sleep disorders additionally may emerge as a consequence of PTSD; thus, there is a bi-directional relationship that, unless treated early, may cause a positive feedback loop, perpetuating a cycle of ever-worsening comorbid PTSD and sleep disorders.
      • Germain A
      • McKeon AB
      • Campbell RL.
      Sleep in PTSD: conceptual model and novel directions in brain-based research and interventions.
      • Laxminarayan S
      • Wang C
      • Ramakrishnan S
      • et al.
      Alterations in sleep electroencephalography synchrony in combat-exposed veterans with post-traumatic stress disorder.
      • Maguire DG
      • Ruddock MW
      • Milanak ME
      • Moore T
      • Cobice D
      • Armour C.
      Sleep, a governor of morbidity in PTSD: a systematic review of biological markers in PTSD-related sleep disturbances.
      • Richards A
      • Kanady JC
      • Neylan TC.
      Sleep disturbance in PTSD and other anxiety-related disorders: an updated review of clinical features, physiological characteristics, and psychological and neurobiological mechanisms.
      Several studies have examined risk factors for PTSD and sleep disorders in military members and found associations with deployment and combat exposure.
      • Gates MA
      • Holowka DW
      • Vasterling JJ
      • Keane TM
      • Marx BP
      • Rosen RC.
      Posttraumatic stress disorder in veterans and military personnel: epidemiology, screening, and case recognition.
      ,
      • Ramchand R
      • Rudavsky R
      • Grant S
      • Tanielian T
      • Jaycox L.
      Prevalence of, risk factors for, and consequences of posttraumatic stress disorder and other mental health problems in military populations deployed to Iraq and Afghanistan.
      ,
      • Foster SN
      • Brock MS
      • Hansen S
      • et al.
      Sleep disorders related to deployment in active duty service members and veterans.
      ,
      • Wright KM
      • Britt TW
      • Bliese PD
      • Adler AB
      • Picchioni D
      • Moore D.
      Insomnia as predictor versus outcome of PTSD and depression among Iraq combat veterans.
      • Caldwell JA
      • Knapik JJ
      • Shing TL
      • Kardouni JR
      • Lieberman HR.
      The association of insomnia and sleep apnea with deployment and combat exposure in the entire population of US army soldiers from 1997 to 2011: a retrospective cohort investigation.
      • Gehrman P
      • Seelig AD
      • Jacobson IG
      • et al.
      Predeployment sleep duration and insomnia symptoms as risk factors for new-onset mental health disorders following military deployment.
      • Wang HE
      • Campbell-Sills L
      • Kessler RC
      • et al.
      Pre-deployment insomnia is associated with post-deployment post-traumatic stress disorder and suicidal ideation in US Army soldiers.
      • Colvonen PJ
      • Almklov E
      • Tripp JC
      • Ulmer CS
      • Pittman JOE
      • Afari N.
      Prevalence rates and correlates of insomnia disorder in post-9/11 veterans enrolling in VA healthcare.
      However, the longitudinal relationships between PTSD and sleep disorders, and the military-related factors that predict their development, are not well established. Determining which military-related factors predict new-onset cases could help in targeting future screening and treatment efforts for PTSD and sleep disorders in military populations.
      OSA and insomnia are 2 of the most prevalent and costly sleep disorders, are often comorbid, and have been associated with PTSD in military populations.
      • Foster SN
      • Brock MS
      • Hansen S
      • et al.
      Sleep disorders related to deployment in active duty service members and veterans.
      ,
      • Mysliwiec V
      • Gill J
      • Lee H
      • et al.
      Sleep disorders in US military personnel: a high rate of comorbid insomnia and obstructive sleep apnea.
      • Moore BA
      • Tison LM
      • Palacios JG
      • Peterson AL
      Mysliwiec V. Incidence of insomnia and obstructive sleep apnea in active duty United States military service members.
      • McLay RN
      • Klam WP
      • Volkert SL.
      Insomnia is the most commonly reported symptom and predicts other symptoms of post-traumatic stress disorder in US service members returning from military deployments.
      • Wright KM
      • Britt TW
      • Bliese PD
      • Adler AB
      • Picchioni D
      • Moore D.
      Insomnia as predictor versus outcome of PTSD and depression among Iraq combat veterans.
      • Caldwell JA
      • Knapik JJ
      • Shing TL
      • Kardouni JR
      • Lieberman HR.
      The association of insomnia and sleep apnea with deployment and combat exposure in the entire population of US army soldiers from 1997 to 2011: a retrospective cohort investigation.
      ,
      • Wang HE
      • Campbell-Sills L
      • Kessler RC
      • et al.
      Pre-deployment insomnia is associated with post-deployment post-traumatic stress disorder and suicidal ideation in US Army soldiers.
      • Colvonen PJ
      • Almklov E
      • Tripp JC
      • Ulmer CS
      • Pittman JOE
      • Afari N.
      Prevalence rates and correlates of insomnia disorder in post-9/11 veterans enrolling in VA healthcare.
      • Colvonen PJ
      • Masino T
      • Drummond SP
      • Myers US
      • Angkaw AC
      • Norman SB.
      Obstructive sleep apnea and posttraumatic stress disorder among OEF/OIF/OND veterans.
      • Rosen RC
      • Cikesh B
      • Fang S
      • et al.
      Posttraumatic stress disorder severity and insomnia-related sleep disturbances: longitudinal associations in a large, gender-balanced cohort of combat-exposed veterans.
      We utilized data from the largest tri-service longitudinal survey of U.S. military members and veterans at 2 time points to address 2 main aims: (1) determine if a bi-directional relationship exists between sleep disorders and PTSD, such that PTSD is associated with the onset of OSA and/or insomnia, and vice versa, and (2) examine military-related factors associated with this potential relationship.

      Participants and methods

      Study population

      Participants were from the Millennium Cohort Study, the largest, longitudinal military study tracking over 200,000 U.S. service members and veterans since 2001.
      • Ryan MA
      • Smith TC
      • Smith B
      • et al.
      Millennium Cohort: enrollment begins a 21-year contribution to understanding the impact of military service.
      Participants were recruited in 4 separate panels between 2001 and 2013, with paper or web-based follow-up surveys occurring every 3-5 years since baseline. The current study uses data from the 2011-2013 (Time 1 [T1]) and 2014-2016 (Time 2 [T2]) Millennium Cohort survey waves to examine the reciprocal, longitudinal relationship between PTSD and OSA and/or insomnia, referred to throughout the rest of the paper as “OSA/insomnia.” This study was approved by the Naval Health Research Center Institutional Review Board and all participants provided informed consent.
      Supplemental data describing the prevalence of OSA and PTSD among the full cohort of participants who responded to the T1 (N = 138,949) or T2 (N = 112,655) surveys are presented in Supplemental Table 1, with specific sample sizes of participants by service status (active duty, Reserve/Guard, or veteran). Sample sizes for insomnia at T1 and T2 in this population have been previously reported.
      • Markwald RR
      • Carey FR
      • Kolaja CA
      • Jacobson IG
      • Cooper AD
      • Chinoy ED.
      Prevalence and predictors of insomnia and sleep medication use in a large tri-service US military sample.
      The base sample for these analyses was drawn from participants who completed both T1 and T2 surveys (N = 98,359). Participants were excluded if they were missing on PTSD at T1 (n = 2029), OSA at T1 (n = 530), insomnia at T1 (n = 2092), PTSD at T2 (n = 8987), OSA at T2 (n = 459), or insomnia at T2 (n = 1830), resulting in an eligible sample of 82,432 participants. Exclusion criteria for the bi-directional relationships examined in these analyses are reported below.

      Analytic Sample

      PTSD at T1 Predicting New-Onset OSA/Insomnia at T2

      Of the eligible sample, 8.4% (n = 6938) of participants had existing OSA at T1 and an additional 11.6% (n = 9579) had existing insomnia at T1 (ie, Insomnia Severity Index [ISI] score ≥15) and were excluded from the population. This resulted in an analytic sample size of 65,915.

      OSA/insomnia at T1 predicting new-onset PTSD at T2

      Of the eligible sample, 13.6% (n = 11,176) of participants screened positive for PTSD and/or reported being diagnosed with PTSD at T1 and were excluded from the population. This resulted in an analytic sample size of 71,256.

      Study measures

      OSA and/or insomnia

      OSA was assessed via a self-reported item asking participants if a doctor had told them they ever had sleep apnea at T1, or in the previous 3 years at T2. Insomnia was assessed as a positive screen using the 7-item ISI, which measures the severity of insomnia symptoms in the last 2 weeks.
      • Bastien CH
      • Vallieres A
      • Morin CM.
      Validation of the Insomnia Severity Index as an outcome measure for insomnia research.
      Items were summed for a total ISI score (range: 0-28) and dichotomized using a cut point of 15, where ≥15 represented a positive screen for insomnia. The cut point of 15 has been established
      • Bastien CH
      • Vallieres A
      • Morin CM.
      Validation of the Insomnia Severity Index as an outcome measure for insomnia research.
      for classifying those with probable cases of “moderate” or “severe” insomnia (vs. no insomnia or only “sub-threshold” insomnia), and was utilized in our previous study on insomnia in the same cohort
      • Markwald RR
      • Carey FR
      • Kolaja CA
      • Jacobson IG
      • Cooper AD
      • Chinoy ED.
      Prevalence and predictors of insomnia and sleep medication use in a large tri-service US military sample.
      and in another recent active duty sample.
      • Taylor DJ
      • Pruiksma KE
      • Hale WJ
      • et al.
      Prevalence, correlates, and predictors of insomnia in the US Army prior to Deployment.
      If participants were missing 1 item on the ISI, this missing item was replaced with an item that was calculated via multiple imputation and summed with the other 6 items to obtain the overall ISI score. Participants missing more than 1 ISI item had their score set to missing and were not included in these analyses.
      OSA and insomnia were combined into a single 4-level item at each time point, classified as having neither OSA nor insomnia, OSA only, insomnia only, or both OSA and insomnia. New-onset OSA/insomnia at T2 was assessed among those who did not have OSA or insomnia at T1.

      PTSD

      A positive screen for current PTSD symptoms was assessed using the 17 item PTSD Checklist-Civilian Version (PCL-C), classified using Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition specific criteria (ie, endorsing moderate or greater on ≥1 intrusion items, ≥2 hyperarousal symptoms, and ≥3 avoidance symptoms).
      • Weathers FW
      • Litz BT
      • Herman DS
      • Huska JA
      • Keane TM.
      The PTSD Checklist (PCL): Reliability, validity, and diagnostic utility.
      ,
      American Psychiatry Association
      Diagnostic and Statistical Manual of Mental Disorders (DSM-IV).
      Participants also self-reported if they had ever been diagnosed with PTSD by a doctor at T1, or in the past 3 years at T2. Participants were categorized using 4 levels: (1) no current PTSD symptoms or self-reported diagnosis, (2) current PTSD symptoms only (positive screen on the PCL-C), (3) self-reported PTSD diagnosis only, or (4) current PTSD symptoms and self-reported diagnosis. New-onset PTSD at T2 was assessed among those who did not have PTSD symptoms or a self-reported diagnosis at T1.

      Other exposures of interest

      Military factors

      Combat deployment was assessed based on Defense Manpower Data Center (DMDC) records of deployments in support of the operations in Iraq and Afghanistan prior to T2 and survey responses to 13 items at T1 and T2 assessing combat-related experiences during deployment (eg, being attacked or ambushed, being wounded or injured), categorized as having never deployed, deployed with no combat (0 combat exposures), deployed with low combat (≤3 combat exposures), or deployed with medium-to-high combat (≥4 combat exposures).
      • Porter B
      • Hoge CW
      • Tobin LE
      • et al.
      Measuring aggregated and specific combat exposures: associations between combat exposure measures and posttraumatic stress disorder, depression, and alcohol-related problems.
      The number of deployments between T1 and T2 were categorized as 0, 1, or ≥2 deployments. The average length of deployment between T1 and T2 was compared with the participant's service branch's average length of deployment (in months: Army=9, Navy=6, Marine Corps=8, and Air Force=4)
      National Research Council
      Returning Home From Iraq and Afghanistan: Assessment of Readjustment Needs of Veterans, Service Members, and Their Families.
      and categorized as at or above branch average (yes or no).
      Military characteristics at T1 (service component, service branch, pay grade, military occupation, and time in service) and separation status prior to T2 were obtained from DMDC records.

      Demographic and behavioral characteristics at T1

      Participants’ age, sex, and race/ethnicity were obtained from DMDC records. Marital status, education level, and all behavioral characteristics were self-reported.
      The mental and physical component summary scores from the Short Form 36 Health Survey for Veterans (SF-36V)
      • Ware J
      • Kosinski M
      • Gandek B.
      SF-36 Health Survey: Manual and Interpretation Guide.
      ,
      • Kazis LE
      • Lee A
      • Spiro 3rd, A
      • et al.
      Measurement comparisons of the medical outcomes study and veterans SF-36 health survey.
      were used as global measures of mental and physical health. Experiences of 5 life stressors in the past 3 years (sexual assault, sexual harassment, physical assault, divorce or separation, major financial problems) were measured using the Social Readjustment Rating Scale,
      • Holmes TH
      • Rahe RH.
      The Social Readjustment Rating Scale.
      based on the documented relationship between stress and mental health and stress and sleep challenges.
      • Leggett A
      • Burgard S
      • Zivin K.
      The impact of sleep disturbance on the association between stressful life events and depressive symptoms.
      Since substance use is associated with sleep issues and mental health,
      • Vandrey R
      • Babson KA
      • Herrmann ES
      • MO Bonn-Miller
      Interactions between disordered sleep, post-traumatic stress disorder, and substance use disorders.
      heavy weekly drinking and smoking status were assessed. Men who reported consuming >14 drinks/week and women who reported consuming >7 drinks/week were considered heavy weekly drinkers
      • Jacobson IG
      • Ryan MA
      • Hooper TI
      • et al.
      Alcohol use and alcohol-related problems before and after military combat deployment.
      ; and tobacco use was categorized as current, former, or never smoker. Finally, self-reported physician-diagnosed traumatic brain injury (TBI) in the past 3 years (not including injuries that had resulted in a concussion only) was examined at T2 based on prior reports of TBI being associated with mental health and sleep.
      • Martindale SL
      • Morissette SB
      • Rowland JA
      • Dolan SL.
      Sleep quality affects cognitive functioning in returning combat veterans beyond combat exposure, PTSD, and mild TBI history.

      Statistical analyses

      Prevalence estimates for T1 and new-onset PTSD and OSA/insomnia and descriptive frequencies of military and demographic characteristics for each analytic sample were reported. Multiple imputation was used among the largest base sample of those responding at T1 and T2 (N = 98,359) to generate 10 datasets with missing covariates imputed. This population was subset to the eligible population (n = 82,432) based on missing outcome variables (OSA, PTSD, insomnia), and then further subset to the analytic new-onset populations at T2 for OSA/insomnia (n = 65,915) and PTSD symptoms/diagnosis (n = 71,256). As it is not recommended to use multiple imputation for outcomes,
      • Kolaja CA
      • Porter B
      • Powell TM
      • Rull RP
      Millennium Cohort Study T
      Multiple imputation validation study: addressing unmeasured survey data in a longitudinal design.
      we did not impute OSA, insomnia (among those missing 2 or more items), or PTSD at the scale level and instead used complete case deletion to address missingness among outcome variables. Among the base population (N = 98,359), the percent missing among covariates prior to multiple imputation ranged between 0 and 7.76%. All statistical models were calculated among each imputed dataset and the reported results were pooled across the 10 imputed datasets.
      • Rubin DB.
      Multiple Imputation for Nonresponse in Surveys.
      Factors associated with attrition in this sample (eg, age, education, sex)
      • Littman AJ
      • Boyko EJ
      • Jacobson IG
      • et al.
      Assessing nonresponse bias at follow-up in a large prospective cohort of relatively young and mobile military service members.
      were included in the multiple imputation model to reduce bias due to missing data.
      • Mostafa T
      • Narayanan M
      • Pongiglione B
      • et al.
      Missing at random assumption made more plausible: evidence from the 1958 British birth cohort.
      Logistic regression models examined whether PTSD at T1, military experiences (combat deployment before T2, separation status at T2, number and length of deployments between T1 and T2), and military characteristics (service branch, pay grade, service component, occupation) predicted new-onset OSA/insomnia at T2. Models were adjusted for all demographic and behavioral characteristics and controlled for time in service and follow-up time. Similarly, additional adjusted logistic regression models examined whether OSA/insomnia at T1, military experiences, and military characteristics predicted new-onset PTSD at T2. All analyses were conducted using SAS 9.4 (SAS Institute Inc., Cary, NC).

      Sensitivity analyses

      To determine the reliability of self-reported OSA, inpatient and outpatient records from the Military Health System Data Repository (MDR) identified potential diagnosed cases of OSA among active duty personnel at T1 who appeared in MDR records at any time between 2000 and 2016 (n = 46,394). Case definitions of OSA within MDR records were classified based on Armed Forces Health Surveillance Branch guidelines.
      Armed Forces Health Surveillance Center
      Obstructive sleep apnea, active component, U.S. Armed Forces, January 2000-December 2009.
      The prevalence of OSA and proportion of observed agreement between self-reported and MDR records were examined. To further determine the level of agreement between both methods, Cohen's kappa coefficient (affected by prevalence) and positive and negative agreement (less affected by prevalence) were calculated at each time point.
      • Cicchetti DV
      • Feinstein AR.
      High agreement but low kappa: II. Resolving the paradoxes.

      Results

      Descriptive statistics: OSA/insomnia and PTSD

      Among participants who did not have OSA/insomnia at T1 (N = 65,915), 2.2% had current PTSD symptoms, 3.1% had a self-reported provider PTSD diagnosis, and 1.1% had current symptoms and a diagnosis. At T2, 5.1% reported new OSA only, 4.1% reported new insomnia only, and 0.9% reported both new OSA and new insomnia (Table 1). Among those who did not have current PTSD symptoms or diagnosis at T1 (N = 71,256), 5.3% reported OSA only, 6.7% screened positive for insomnia only, and 1.3% had comorbid OSA and insomnia. At T2, 3.9% developed PTSD symptoms, 2.9% newly reported provider-diagnosed PTSD, and 2.1% newly reported symptoms and a diagnosis (Table 1).
      Table 1Descriptive frequencies for PTSD and OSA and/or insomnia symptoms among Millennium Cohort Study participants for each analytic population
      Analytic sample: PTSD at T1 predicting OSA/insomnia at T2, N = 65,915Analytic sample: OSA/insomnia at T1 predicting PTSD at T2, N = 71,256
      Prevalence of PTSD and OSA/insomniaPTSD at T1New onset OSA/insomnia at T2OSA/insomnia at T1New onset PTSD at T2
      n (%)n (%)n (%)n (%)
      PTSD
      No symptoms or self-reported diagnosis61,699 (93.6)64,987 (91.2)
      Current symptoms (PCL) only1419 (2.2)2740 (3.9)
      Diagnosis (self-reported) only2056 (3.1)2052 (2.9)
      Current symptoms and diagnosis741 (1.1)1477 (2.1)
      OSA and/or insomnia
      No OSA or insomnia59,254 (89.9)61,699 (86.6)
      OSA only3382 (5.1)3808 (5.3)
      Insomnia only2686 (4.1)4803 (6.7)
      OSA and insomnia593 (0.9)946 (1.3)
      OSA, obstructive sleep apnea; PCL, PTSD Checklist-Civilian Version; PTSD, post-traumatic stress disorder; T1, Time 1 (2011-2013); T2, Time 2 (2014-2016).

      Descriptive statistics: Military characteristics and demographics

      The sample for these 2 populations were similar in terms of military experiences and characteristics, as well as demographics (Table 2). Over half the sample deployed before T1 and were separated from military service before the end of the study, and most did not deploy between T1 and T2. The majority of both samples were active duty (vs. Reserve/Guard), serving in the Army or Air Force, enlisted members, and serving in occupations other than combat specialist, functional support/administrative, or healthcare. This sample was predominantly younger (<35 years), male, non-Hispanic White and married (Table 2). The mean length of follow-up for participants between T1 and T2 was 3.4 years (SD: 0.6, range: 1.0-5.4).
      Table 2Descriptive frequencies for military experiences and characteristics and demographics among Millennium Cohort Study participants for each analytic population
      Analytic sample: PTSD at T1 predicting OSA/insomnia at T2, N = 65,915Analytic sample: OSA/insomnia at T1 predicting PTSD at T2, N = 71,256
      Military experiences and characteristics%%
      Combat deployment before T1
       Never deployed41.643.2
       Deployed, no combat33.132.6
       Deployed, low combat12.011.8
       Deployed, medium-to-high combat13.212.4
      Separation status
       On active service at T248.146.5
       Separated before T138.840.2
       Separated between T1 and T213.113.4
      No. of deployments, T1-T2
       080.681.2
       113.012.6
       2 or more6.46.2
      Deployed equal to or longer than branch average from T1 to T2
      Average deployment lengths for each branch: Army = 9 months, Navy/Coast Guard = 6 months, Marine Corps = 8 months, Air Force = 4 months. No (less than branch average) category also includes individuals who did not deploy during the time frame.
       No94.194.2
       Yes5.95.8
      Service component
       Reserve/Guard45.644.7
       Active duty54.455.3
      Service branch
       Army41.841.4
       Navy/Coast Guard18.619.1
       Marine Corps7.37.0
       Air Force32.332.4
      Pay grade
       Enlisted69.069.4
       Officer31.030.6
      Occupation
       Combat specialist18.317.9
       Functional support/admin17.918.2
       Healthcare13.212.8
       Other50.751.1
      Demographic characteristics at T1%%
      Age, y
       <256.96.7
       25-3445.443.0
       35-4421.121.3
       45+26.629.0
      Sex
       Male70.371.0
       Female29.829.0
      Race/ethnicity
       White78.578.0
       Black8.59.0
       Other12.913.0
      Marital status
       Single20.119.3
       Married67.168.0
       Separated/divorced/widowed12.812.8
      Education level
       Associates degree or less51.651.9
       Bachelor's degree or higher48.448.2
      OSA, obstructive sleep apnea; PTSD, post-traumatic stress disorder; T1, Time 1 (2011-2013); T2, Time 2 (2014-2016).
      All military and demographic characteristics are assessed at T1 unless otherwise specified.
      a Average deployment lengths for each branch: Army = 9 months, Navy/Coast Guard = 6 months, Marine Corps = 8 months, Air Force = 4 months. No (less than branch average) category also includes individuals who did not deploy during the time frame.

      Adjusted models: PTSD at T1 predicting new-onset OSA/insomnia at T2

      In adjusted models examining whether PTSD at T1 predicted new-onset OSA and/or insomnia at T2 (Table 3), participants with a PTSD diagnosis at T1 were significantly more likely to develop OSA only or both OSA and insomnia at T2 compared with those who did not have PTSD symptoms or diagnosis at T1. Those with current PTSD symptoms, a diagnosis, or both were more likely to develop insomnia without OSA compared with individuals without PTSD symptoms or diagnosis at T1. Being deployed with medium-to-high combat exposure (vs. never being deployed), separating from service between T1 and T2 (vs. remaining on active service), and serving in the Army (vs. Air Force) predicted new-onset OSA only, insomnia only, or comorbid conditions at T2. Being an officer (vs. enlisted) was protective against the development of these 3 outcomes. Factors that predicted new-onset OSA only included deployment without combat (vs. not deploying) and serving in non-combat occupations (eg functional support/administration and health care). Factors that predicted new-onset insomnia only included serving in the Navy or Marine Corps (vs. Air Force). Reserve/Guard component (vs. active duty) was protective for new-onset OSA only and insomnia only. Adjusted associations between demographic characteristics at T1 and new-onset OSA and/or insomnia at T2 are presented in Supplemental Table 2.
      Table 3Adjusted associations for new-onset OSA and insomnia symptoms at T1 among Millennium Cohort Study participants without OSA or insomnia at T1, N = 65,915
      Variables at T1
      All variables are assessed at T1 unless otherwise specified.
      Adjusted ORs of new-onset OSA/insomnia at T2 (ref: Neither)
      New-onset OSA only, n = 3382New-onset insomnia only, n = 2686New-onset OSA and insomnia, n = 593
      OR (95% CI)OR (95% CI)OR (95% CI)
      PTSD
      PTSD at T1
       No symptoms or diagnosisRefRefRef
       Symptoms (PCL) only1.09 (0.85-1.39)1.27 (1.06-1.54)1.34 (0.90-1.99)
       Diagnosis (self-reported) only1.50 (1.26-1.79)1.71 (1.46-2.01)1.67 (1.21-2.30)
       Symptoms and diagnosis1.26 (0.94-1.70)1.41 (1.12-1.78)1.39 (0.88-2.19)
      Military experiences
      Combat deployment before T1
       Never deployedRefRefRef
       Deployed, no combat1.14 (1.04-1.25)0.94 (0.85-1.04)1.02 (0.81-1.27)
       Deployed, low combat1.11 (0.97-1.26)1.02 (0.89-1.18)1.20 (0.89-1.62)
       Deployed, medium-to-high combat1.26 (1.11-1.43)1.40 (1.24-1.59)1.62 (1.25-2.11)
      Separation status
       On active service at T2RefRefRef
       Separated before T10.88 (0.79-0.99)1.01 (0.90-1.13)0.77 (0.59-1.00)
       Separated between T1 and T22.41 (2.18-2.66)1.61 (1.44-1.80)3.18 (2.59-3.92)
      No. of deployments, T1-T2
       0RefRefRef
       11.03 (0.89-1.19)1.06 (0.91-1.24)1.00 (0.73-1.37)
       2 or more0.94 (0.79-1.12)1.13 (0.93-1.38)0.90 (0.59-1.39)
      Deployed equal to or longer branch average from T1 to T2
      Average deployment lengths for each branch: Army = 9 months, Navy/Coast Guard = 6 months, Marine Corps = 8 months, Air Force = 4 months. No (less than branch average) category also includes individuals who did not deploy during the time frame.
       No0.93 (0.77-1.12)0.78 (0.64-0.96)0.76 (0.51-1.15)
       YesRefRefRef
      Military characteristics
      Service component
       Reserve/Guard0.66 (0.61-0.72)0.89 (0.82-0.98)0.87 (0.73-1.05)
       Active dutyRefRefRef
      Service branch
       Army1.12 (1.02-1.22)1.44 (1.29-1.61)1.56 (1.24-1.96)
       Navy/Coast Guard1.07 (0.96-1.19)1.28 (1.12-1.46)1.30 (0.98-1.72)
       Marine Corps0.96 (0.81-1.14)1.63 (1.38-1.91)1.18 (0.81-1.72)
       Air ForceRefRefRef
      Pay grade
       EnlistedRefRefRef
       Officer0.77 (0.69-0.85)0.60 (0.52-0.69)0.63 (0.48-0.83)
      Occupation
       Combat specialistRefRefRef
       Functional support/admin1.28 (1.13-1.44)1.04 (0.90-1.21)1.16 (0.88-1.54)
       Healthcare1.38 (1.20-1.59)1.00 (0.85-1.17)0.92 (0.66-1.30)
       Other1.17 (1.06-1.30)1.10 (0.97-1.24)1.01 (0.80-1.28)
      CI, confidence interval; OR, odds ratio; OSA, obstructive sleep apnea; PCL, PTSD Checklist-Civilian Version; PTSD, post-traumatic stress disorder; T1, Time 1 (2011-2013); T2, Time 2 (2014-2016).
      Logistic regression model is further adjusted for age, sex, race/ethnicity, marital status, education level, heavy weekly drinking, smoking status, life stressors, diagnosed traumatic brain injury at T2, mental component score, physical component score, time in service at T1, and follow-up time between T1 and T2. All ORs and CIs represent pooled estimates across multiple imputation models.
      Bolded values are statistically significant (p < .05).
      a All variables are assessed at T1 unless otherwise specified.
      b Average deployment lengths for each branch: Army = 9 months, Navy/Coast Guard = 6 months, Marine Corps = 8 months, Air Force = 4 months. No (less than branch average) category also includes individuals who did not deploy during the time frame.

      Adjusted models: OSA/insomnia at T1 predicting new-onset PTSD at T2

      In adjusted models examining whether OSA and/or insomnia at T1 predicted new-onset PTSD at T2 (Table 4), participants with OSA only, insomnia only, and both OSA and insomnia were significantly more likely to develop PTSD symptoms, newly report a diagnosis, or both compared with those who had neither OSA nor insomnia at T1. One exception was that the measure of effect for those with insomnia only predicting a PTSD diagnosis was not statistically significant. Being deployed with or without combat exposure before T2 (vs. never deploying), separating from service before T1 or between T1 and T2 (vs. remaining on active service), and serving in the Army, Navy, or Marine Corps (vs. Air Force; except Navy was not statistically significant for newly reported PTSD diagnosis only) predicted new-onset PTSD symptoms, diagnosis, or both symptoms and diagnosis at T2. Being an officer (vs. enlisted) was protective against the development of these 3 outcomes. Factors that predicted newly reporting a PTSD diagnosis only included deploying either 1 or ≥2 times (vs. 0) and Reserve/Guard component (vs. active duty) (Table 4). Adjusted associations between demographic characteristics at T1 and new-onset PTSD at T2 are presented in Supplemental Table 3.
      Table 4Adjusted associations for new-onset PTSD at T2 among Millennium Cohort Study participants without PTSD at T1 (N = 71,256)
      Variables
      All variables are assessed at T1 unless otherwise specified.
      New-onset PTSD at T2 (ref: No symptoms or diagnosis)
      Symptoms (PCL) only, n = 2740Diagnosis (self-reported) only, n = 2052Symptoms and diagnosis, n = 1477
      OR (95% CI)OR (95% CI)OR (95% CI)
      OSA and insomnia
      OSA/insomnia at T1
       NeitherRefRefRef
       OSA only1.23 (1.02-1.47)1.68 (1.40-2.01)1.29 (1.01-1.65)
       Insomnia only1.60 (1.43-1.79)1.13 (0.96-1.32)1.59 (1.36-1.86)
       OSA and insomnia1.31 (1.03-1.67)1.56 (1.17-2.09)1.84 (1.37-2.47)
      Military experiences
      Combat deployment before T1
       Never deployedRefRefRef
       Deployed, no combat1.15 (1.04-1.27)1.76 (1.55-2.00)1.49 (1.28-1.73)
       Deployed, low combat1.04 (0.91-1.21)2.28 (1.95-2.68)1.46 (1.20-1.79)
       Deployed, medium-to-high combat1.54 (1.36-1.76)4.31 (3.74-4.97)3.54 (3.01-4.17)
      Separation status
       On active service at T2RefRefRef
       Separated before T11.17 (1.04-1.31)1.29 (1.13-1.48)1.33 (1.12-1.57)
       Separated between T1 and T21.76 (1.57-1.97)2.66 (2.36-2.99)3.29 (2.86-3.78)
      No. of deployments, T1-T2
       0RefRefRef
       11.14 (0.97-1.33)1.40 (1.19-1.64)1.42 (1.17-1.72)
       2 or more0.86 (0.69-1.07)1.38 (1.12-1.70)1.21 (0.91-1.59)
      Deployed equal to or longer than branch average from T1 to T2
      Average deployment lengths for each branch: Army = 9 months, Navy/Coast Guard = 6 months, Marine Corps = 8 months, Air Force = 4 months. No (less than branch average) category also includes individuals who did not deploy during the time frame.
       No1.19 (0.95-1.49)0.94 (0.76-1.17)1.00 (0.77-1.29)
       YesRefRefRef
      Military characteristics
      Service component
       Reserve/Guard1.00 (0.91-1.09)1.21 (1.09-1.34)1.09 (0.96-1.23)
       Active dutyRefRefRef
      Service branch
       Army1.43 (1.28-1.59)1.82 (1.61-2.05)1.87 (1.60-2.18)
       Navy/Coast Guard1.23 (1.08-1.40)1.14 (0.97-1.34)1.27 (1.04-1.55)
       Marine Corps1.55 (1.32-1.82)1.70 (1.40-2.07)1.64 (1.30-2.08)
       Air ForceRefRefRef
      Pay grade
       EnlistedRefRefRef
       Officer0.66 (0.57-0.76)0.66 (0.56-0.76)0.53 (0.44-0.65)
      Occupation
       Combat specialistRefRefRef
       Functional support/admin0.93 (0.80-1.08)0.82 (0.70-0.96)0.85 (0.70-1.04)
       Healthcare0.95 (0.81-1.12)1.14 (0.96-1.35)0.99 (0.80-1.23)
       Other1.00 (0.89-1.13)0.88 (0.77-1.00)0.95 (0.81-1.11)
      CI, confidence interval; OR, odds ratio; OSA, obstructive sleep apnea; PCL, PTSD Checklist-Civilian Version; PTSD, post-traumatic stress disorder; T1, Time 1 (2011-2013), T2: Time 2 (2014-2016).
      Logistic regression model is further adjusted for age, sex, race/ethnicity, marital status, education level, heavy weekly drinking, smoking status, life stressors, diagnosed traumatic brain injury at T2, mental component score, physical component score, time in service at T1, and follow-up time between T1 and T2. All ORs and CIs represent pooled estimates across multiple imputation models.
      Bolded values are statistically significant (p < .05).
      a All variables are assessed at T1 unless otherwise specified.
      b Average deployment lengths for each branch: Army = 9 months, Navy/Coast Guard = 6 months, Marine Corps = 8 months, Air Force = 4 months. No (less than branch average) category also includes individuals who did not deploy during the time frame.

      Sensitivity analyses for OSA classifications

      The prevalence and agreement of self-reported and MDR-classified OSA are reported in Table 5. At both T1 and T2, the prevalence of OSA via MDR records was lower than the self-reported prevalence, though the proportion of observed agreement between the methods was high (>90%). The kappa coefficients indicated a substantial level of reliability
      • Landis JR
      • Koch GG.
      The measurement of observer agreement for categorical data.
      (κ > 0.61) between both methods at both time points, and the proportions of specific agreement indicated moderate positive agreement and good negative agreement.
      • Cicchetti DV
      • Feinstein AR.
      High agreement but low kappa: II. Resolving the paradoxes.
      Table 5Prevalence and agreement of self-reported and MDR-classified OSA among active duty Millennium Cohort Study participants, T1 and T2 (n = 46,394)
      Time pointPrevalenceProportion of observed agreementKappa statisticsProportions of specific agreement
      Self-reportMDRκ95% CIPositive agreementNegative agreement
      OSA at T19.1%7.4%94.8%0.660.65-0.6768.9%97.2%
      OSA at T214.6%13.3%92.8%0.700.69-0.7174.3%95.9%
      CI, confidence interval; MDR, Military Health System Data Repository; OR, odds ratio; OSA, obstructive sleep apnea; T1, Time 1 (2011-2013); T2, Time 2 (2014-2016).

      Discussion

      In this study, we utilized the largest longitudinal U.S. military cohort to examine whether military-related factors and PTSD predict OSA and/or insomnia at follow-up 3-5 years later, and vice versa. Findings suggested that there were strong bi-directional relationships between new-onset PTSD symptoms and/or diagnosis and OSA and/or insomnia, which was sometimes revealed by health care utilization. In addition, several military-related factors increased the odds for new-onset cases of both conditions. Specifically, prior deployment (especially with combat exposures), recent separation from the military, and service branch (Army, Navy, Marine Corps) predicted new-onset cases, while higher military rank was protective. Supplemental findings revealed that the prevalence of PTSD among the full Millennium Cohort Study population was higher among those with OSA/insomnia than those without sleep disorders, and vice versa; both PTSD and OSA/insomnia were higher among veterans than active duty or Reserve/Guard members.
      The longitudinal analyses showed the strength of associations between PTSD symptoms and/or diagnosis and OSA/insomnia, revealing significant bi-directional relationships. Specifically, only self-reported provider-diagnosed PTSD without current symptoms was associated with greater likelihood of newly reporting OSA or comorbid OSA and insomnia, while having PTSD symptoms, a diagnosis, or both were all associated with new-onset insomnia. This may reflect differences in healthcare-seeking or utilization behavior since OSA and PTSD diagnosis would be reported by those who have received care, while screening positive for PTSD and insomnia symptoms may reflect present conditions among those who do not wish or have yet to seek care. When examining OSA/insomnia predicting PTSD, both OSA and insomnia were associated with the development of PTSD symptoms and/or diagnosis, but having insomnia only was not significantly associated with a newly reported PTSD diagnosis. This may reflect that some individuals with insomnia symptoms alone may not necessarily utilize health care services where they would receive a provider PTSD diagnosis, and also that PTSD is a complex condition with many potential causes that are difficult to disentangle. Prior studies indicate that sleep disorders are comorbid with or increase risk for PTSD.
      • Foster SN
      • Brock MS
      • Hansen S
      • et al.
      Sleep disorders related to deployment in active duty service members and veterans.
      ,
      • Mysliwiec V
      • Gill J
      • Lee H
      • et al.
      Sleep disorders in US military personnel: a high rate of comorbid insomnia and obstructive sleep apnea.
      ,
      • McLay RN
      • Klam WP
      • Volkert SL.
      Insomnia is the most commonly reported symptom and predicts other symptoms of post-traumatic stress disorder in US service members returning from military deployments.
      • Wright KM
      • Britt TW
      • Bliese PD
      • Adler AB
      • Picchioni D
      • Moore D.
      Insomnia as predictor versus outcome of PTSD and depression among Iraq combat veterans.
      • Caldwell JA
      • Knapik JJ
      • Shing TL
      • Kardouni JR
      • Lieberman HR.
      The association of insomnia and sleep apnea with deployment and combat exposure in the entire population of US army soldiers from 1997 to 2011: a retrospective cohort investigation.
      • Gehrman P
      • Seelig AD
      • Jacobson IG
      • et al.
      Predeployment sleep duration and insomnia symptoms as risk factors for new-onset mental health disorders following military deployment.
      • Wang HE
      • Campbell-Sills L
      • Kessler RC
      • et al.
      Pre-deployment insomnia is associated with post-deployment post-traumatic stress disorder and suicidal ideation in US Army soldiers.
      • Colvonen PJ
      • Almklov E
      • Tripp JC
      • Ulmer CS
      • Pittman JOE
      • Afari N.
      Prevalence rates and correlates of insomnia disorder in post-9/11 veterans enrolling in VA healthcare.
      • Colvonen PJ
      • Masino T
      • Drummond SP
      • Myers US
      • Angkaw AC
      • Norman SB.
      Obstructive sleep apnea and posttraumatic stress disorder among OEF/OIF/OND veterans.
      • Rosen RC
      • Cikesh B
      • Fang S
      • et al.
      Posttraumatic stress disorder severity and insomnia-related sleep disturbances: longitudinal associations in a large, gender-balanced cohort of combat-exposed veterans.
      Sleep regulates many key functions in the brain, thus, sleep disorders can cause dysregulated neurophysiology and behavior that may heighten PTSD risk if exposed to significant stressors, and vice versa.
      • Germain A
      • McKeon AB
      • Campbell RL.
      Sleep in PTSD: conceptual model and novel directions in brain-based research and interventions.
      • Laxminarayan S
      • Wang C
      • Ramakrishnan S
      • et al.
      Alterations in sleep electroencephalography synchrony in combat-exposed veterans with post-traumatic stress disorder.
      • Maguire DG
      • Ruddock MW
      • Milanak ME
      • Moore T
      • Cobice D
      • Armour C.
      Sleep, a governor of morbidity in PTSD: a systematic review of biological markers in PTSD-related sleep disturbances.
      • Richards A
      • Kanady JC
      • Neylan TC.
      Sleep disturbance in PTSD and other anxiety-related disorders: an updated review of clinical features, physiological characteristics, and psychological and neurobiological mechanisms.
      Despite only examining these associations between 2 surveys 3-5 years apart, our longitudinal findings provide support for bi-directional associations between PTSD and sleep disorders. Studies with more frequent surveys and/or longer follow-up time periods are needed to further explore these associations.
      We also examined which military-related factors were associated with new-onset PTSD or OSA/insomnia. Deployment (especially with medium-to-high combat exposure) was a significant risk factor. Deployment conditions can be stressful and, when combined with combat exposures, may be traumatic for some individuals, as previous studies have also reported associations of PTSD, OSA/insomnia, and/or other sleep issues with deployment.
      • Gates MA
      • Holowka DW
      • Vasterling JJ
      • Keane TM
      • Marx BP
      • Rosen RC.
      Posttraumatic stress disorder in veterans and military personnel: epidemiology, screening, and case recognition.
      ,
      • Ramchand R
      • Rudavsky R
      • Grant S
      • Tanielian T
      • Jaycox L.
      Prevalence of, risk factors for, and consequences of posttraumatic stress disorder and other mental health problems in military populations deployed to Iraq and Afghanistan.
      ,
      • Foster SN
      • Brock MS
      • Hansen S
      • et al.
      Sleep disorders related to deployment in active duty service members and veterans.
      ,
      • McLay RN
      • Klam WP
      • Volkert SL.
      Insomnia is the most commonly reported symptom and predicts other symptoms of post-traumatic stress disorder in US service members returning from military deployments.
      ,
      • Caldwell JA
      • Knapik JJ
      • Shing TL
      • Kardouni JR
      • Lieberman HR.
      The association of insomnia and sleep apnea with deployment and combat exposure in the entire population of US army soldiers from 1997 to 2011: a retrospective cohort investigation.
      • Gehrman P
      • Seelig AD
      • Jacobson IG
      • et al.
      Predeployment sleep duration and insomnia symptoms as risk factors for new-onset mental health disorders following military deployment.
      • Wang HE
      • Campbell-Sills L
      • Kessler RC
      • et al.
      Pre-deployment insomnia is associated with post-deployment post-traumatic stress disorder and suicidal ideation in US Army soldiers.
      • Colvonen PJ
      • Almklov E
      • Tripp JC
      • Ulmer CS
      • Pittman JOE
      • Afari N.
      Prevalence rates and correlates of insomnia disorder in post-9/11 veterans enrolling in VA healthcare.
      Interestingly, except for increased odds at T2 of new-onset OSA among those having prior deployment with no combat exposure, prior deployment with either no or low combat was not associated with new-onset OSA/insomnia. However, there was increased risk for new-onset PTSD symptoms and/or diagnosis with any level of combat exposure in prior deployments, or with 1 or multiple deployments between T1 and T2 (for new-onset diagnosis or both symptoms and diagnosis). Thus, deployment with medium-to-high combat exposure is a strong risk factor for new-onset OSA/insomnia and PTSD, but lower levels of combat appear to confer risk for only new-onset PTSD. These findings are supportive of targeted screening for sleep issues and PTSD among individuals deploying in conditions with high levels of combat exposure.
      Recent separation from the military was another strong risk factor for both new-onset OSA/insomnia and PTSD. Previous Millennium Cohort studies
      • Markwald RR
      • Carey FR
      • Kolaja CA
      • Jacobson IG
      • Cooper AD
      • Chinoy ED.
      Prevalence and predictors of insomnia and sleep medication use in a large tri-service US military sample.
      ,
      • Cooper AD
      • Kolaja CA
      • Markwald RR
      • Jacobson IG
      • Chinoy ED.
      Longitudinal associations of military-related factors on self-reported sleep among U.S. service members.
      also found that recent separation was a risk factor for reported trouble sleeping and new-onset insomnia (using ISI), regardless of prior PTSD status. Thus, our findings indicate a risk of developing sleep disorders after separation, and sleep health promotion interventions should be utilized, especially when military members transition back to civilian life.
      Service branch was another risk factor. While service in any military branch presents stressors and sleep challenges, prior studies typically found that PTSD
      • Judkins JL
      • Moore BA
      • Collette TL
      • Hale WJ
      • Peterson AL
      • Morissette SB.
      Incidence rates of posttraumatic stress disorder over a 17-year period in active duty military service members.
      ,
      • Ramchand R
      • Rudavsky R
      • Grant S
      • Tanielian T
      • Jaycox L.
      Prevalence of, risk factors for, and consequences of posttraumatic stress disorder and other mental health problems in military populations deployed to Iraq and Afghanistan.
      and sleep issues
      • Moore BA
      • Tison LM
      • Palacios JG
      • Peterson AL
      Mysliwiec V. Incidence of insomnia and obstructive sleep apnea in active duty United States military service members.
      ,
      • Markwald RR
      • Carey FR
      • Kolaja CA
      • Jacobson IG
      • Cooper AD
      • Chinoy ED.
      Prevalence and predictors of insomnia and sleep medication use in a large tri-service US military sample.
      ,
      • Cooper AD
      • Kolaja CA
      • Markwald RR
      • Jacobson IG
      • Chinoy ED.
      Longitudinal associations of military-related factors on self-reported sleep among U.S. service members.
      were greatest in the Army and/or Marine Corps, with mixed results for Navy. Thus, while our findings support the associations between OSA/insomnia and PTSD in Army, Navy, and Marine Corps personnel, differing requirements, experiences, combat exposures, and cultures may confer different health risk profiles among the service branches.
      The only protective factor we found for both new-onset PTSD and OSA/insomnia was being an officer, compared with enlisted. This also aligns with previous Millennium Cohort findings that officers are protected from other new-onset sleep issues (eg, short sleep duration, insomnia, sleep medication use).
      • Markwald RR
      • Carey FR
      • Kolaja CA
      • Jacobson IG
      • Cooper AD
      • Chinoy ED.
      Prevalence and predictors of insomnia and sleep medication use in a large tri-service US military sample.
      ,
      • Cooper AD
      • Kolaja CA
      • Markwald RR
      • Jacobson IG
      • Chinoy ED.
      Longitudinal associations of military-related factors on self-reported sleep among U.S. service members.
      Thus, higher military rank may confer benefits to sleep and mental health risks. Two factors interestingly showed a mix of risk and protection: being in the Reserve/Guard and separating prior to T1 both increased risk for new-onset PTSD but were protective for new-onset OSA/insomnia. This contrasts with recent separation (ie, between T1 and T2), which was a risk factor for both. These complex findings indicate that the risk for new-onset PTSD is high around the transition period, but those separated from service for longer times are at lower risk after passing through the higher-risk transition period. This may be due in part to service members reporting health issues immediately prior to transition out of the military, even when the condition may have existed earlier but was not reported. Thus, separation has a complex risk timeline that should be further examined. Additionally, all occupations (compared with combat specialist) had increased risk for new-onset OSA only, but being in a functional support/administrative occupation was protective for newly diagnosed PTSD only. Our previous reports
      • Markwald RR
      • Carey FR
      • Kolaja CA
      • Jacobson IG
      • Cooper AD
      • Chinoy ED.
      Prevalence and predictors of insomnia and sleep medication use in a large tri-service US military sample.
      ,
      • Cooper AD
      • Kolaja CA
      • Markwald RR
      • Jacobson IG
      • Chinoy ED.
      Longitudinal associations of military-related factors on self-reported sleep among U.S. service members.
      showed mixed results for military occupations, wherein we found no significant associations for new-onset short sleep duration or insomnia, and only healthcare occupations increased risk for new-onset trouble sleeping and sleep medication use.
      There are limitations to this study. First, we used self-reported survey data, which are subject to reporting and recall biases and do not always represent a clinical diagnosis of the outcomes under study. As 2 time points were utilized for the analyses, participants were only eligible if they responded at both points, which may bias the sample. This limitation was alleviated somewhat with the application of multiple imputation models that included characteristics associated with attrition.
      • Mostafa T
      • Narayanan M
      • Pongiglione B
      • et al.
      Missing at random assumption made more plausible: evidence from the 1958 British birth cohort.
      Also, the ISI cutoff of ≥15 has been utilized in several previous military samples
      • Markwald RR
      • Carey FR
      • Kolaja CA
      • Jacobson IG
      • Cooper AD
      • Chinoy ED.
      Prevalence and predictors of insomnia and sleep medication use in a large tri-service US military sample.
      ,
      • Taylor DJ
      • Pruiksma KE
      • Hale WJ
      • et al.
      Prevalence, correlates, and predictors of insomnia in the US Army prior to Deployment.
      and is more conservative than lower cutoffs sometimes used with general population samples
      • Bastien CH
      • Vallieres A
      • Morin CM.
      Validation of the Insomnia Severity Index as an outcome measure for insomnia research.
      that include “sub-threshold” insomnia. Thus, insomnia cases may be underestimated in the current study. Formal validation of the ISI in military samples is needed. Furthermore, OSA and PTSD may be underestimated in this study given that both OSA and PTSD tend to be underdiagnosed in medical settings, and participants actively serving may be motivated to under-report PTSD symptoms. Therefore, even by casting a wider net that included the assessment of both PTSD symptoms on the PCL and self-reported provider-diagnosed PTSD, this condition may still be underrepresented in this sample. In addition, the presence of PTSD, OSA, and insomnia were not assessed prior to T1 and some participants who screened negative for these at T1 may have a previous history that was not accounted for. Likewise, the screening tools for insomnia and PTSD assessed symptoms over the past 2 weeks and month, respectively, so the prevalence of these conditions at T2 may not represent symptomology present between T1 and T2. Also, we were unable to assess possible treatment for PTSD, OSA, or insomnia prior to T1 and/or between T1 and T2. In addition, we were unable to evaluate TBI at T1 since this item was not assessed until the T2 survey, which means we could not disentangle the temporal sequence of TBI in relation to sleep disorders and PTSD. Lastly, we were unable to assess fitness for duty, which may potentially deter self-reporting PTSD or sleep disorders prior to separation from service.
      Despite these limitations, a significant strength was the use of a large, representative sample of service members and veterans from all service branches and components. Additionally, longitudinal data allowed us to establish temporality in the development of the outcomes of interest in relation to baseline conditions. Moreover, the measures used for insomnia and PTSD symptoms are derived from widely used and validated measures,
      • Bastien CH
      • Vallieres A
      • Morin CM.
      Validation of the Insomnia Severity Index as an outcome measure for insomnia research.
      lending credibility to the classifications of each outcome. While OSA was assessed through a single self-report item, sensitivity analyses were used to validate this self-report item with OSA diagnoses in medical records among a subpopulation of active duty participants. Though these analyses suggested there was a small potential over-reporting of OSA via self-report compared with medical records, the overall reliability between both methods was good. Results also indicated that while there was higher agreement between both methods in determining negative OSA diagnoses than in determining positive diagnoses, the self-report item provided adequate coverage in capturing the presence of OSA, strengthening our ability to draw conclusions based on this self-reported measure.

      Conclusions

      In a large, representative U.S. military cohort assessed at 2 time points, 3-5 years apart, longitudinal analyses showed that having either PTSD or OSA/insomnia was bi-directionally predictive for new onset of the other condition, and several military-related factors increased risk, including prior deployment (especially with medium-to-high combat exposure), recent separation, and serving in the Army, Navy, or Marine Corps, while being an officer was a protective factor. Furthermore, the prevalence of PTSD and OSA/insomnia was high, often comorbid, and greatest among veterans. Utilizing the identified risk factors from this study, sleep and mental health screenings and interventions can be better targeted to the military communities and individuals at highest risk for developing PTSD and sleep disorders.

      Declaration of conflict of interest

      The authors have no conflict of interest to declare.

      Disclaimer

      Rachel R. Markwald is an employee of the U.S. Government. This work was prepared as part of my official duties. Title 17, U.S.C. §105 provides that copyright protection under this title is not available for any work of the U.S. Government. Title 17, U.S.C. §101 defines a U.S. Government work as work prepared by a military service member or employee of the U.S. Government as part of that person's official duties. Report No. 21-34 was supported by the Military Operational Medicine Research Program, Defense Health Program, and the Department of Veterans Affairs under work unit no. 60002. The views expressed in this article are those of the authors and do not necessarily reflect the official policy or position of the Department of the Navy, Department of Defense, nor the U.S. Government. The study protocol was approved by the Naval Health Research Center Institutional Review Board in compliance with all applicable Federal regulations governing the protection of human subjects. Research data were derived from an approved Naval Health Research Center Institutional Review Board protocol, number NHRC.2000.0007.

      Funding

      This research was supported by funding from the Military Operational Medicine Research Program , Defense Health Program , and the Department of Veterans Affairs .

      Acknowledgments

      In addition to the authors, the Millennium Cohort Study team includes Jennifer Belding, PhD; Satbir Boparai, MBA; Ania Bukowinski, MPH; Sheila Castañeda, PhD; Clinton Hall, MPH, PhD; Toni Rose Geronimo-Hara, MPH; William Lee; Rayna Matsuno, PhD; Deanne Millard; Chiping Nieh, PhD; Ben Porter, PhD; Anna Rivera, MPH; Teresa Powell, MS; Rudolph Rull, PhD; Beverly Sheppard; Daniel Trone, PhD; Jennifer Walstrom; and Steven Warner, MPH. The authors also appreciate contributions from the Deployment Health Research Department, Millennium Cohort Family Study, Birth and Infant Health Research Team, and Henry M. Jackson Foundation. We greatly appreciate the contributions of the Millennium Cohort Study participants.

      Appendix. Supplementary materials

      References

        • Gates MA
        • Holowka DW
        • Vasterling JJ
        • Keane TM
        • Marx BP
        • Rosen RC.
        Posttraumatic stress disorder in veterans and military personnel: epidemiology, screening, and case recognition.
        Psychol Serv. 2012; 9: 361-382
        • Walter KH
        • Levine JA
        • Highfill-McRoy RM
        • Navarro M
        • Thomsen CJ.
        Prevalence of posttraumatic stress disorder and psychological comorbidities among U.S. active duty service members, 2006-2013.
        J Trauma Stress. 2018; 31: 837-844
        • Judkins JL
        • Moore BA
        • Collette TL
        • Hale WJ
        • Peterson AL
        • Morissette SB.
        Incidence rates of posttraumatic stress disorder over a 17-year period in active duty military service members.
        J Trauma Stress. 2020; 33: 994-1006
        • Ramchand R
        • Rudavsky R
        • Grant S
        • Tanielian T
        • Jaycox L.
        Prevalence of, risk factors for, and consequences of posttraumatic stress disorder and other mental health problems in military populations deployed to Iraq and Afghanistan.
        Curr Psychiatry Rep. 2015; 17: 37
        • Germain A
        • McKeon AB
        • Campbell RL.
        Sleep in PTSD: conceptual model and novel directions in brain-based research and interventions.
        Curr Opin Psychol. 2017; 14: 84-89
        • Laxminarayan S
        • Wang C
        • Ramakrishnan S
        • et al.
        Alterations in sleep electroencephalography synchrony in combat-exposed veterans with post-traumatic stress disorder.
        Sleep. 2020; 43: zsaa006
        • Maguire DG
        • Ruddock MW
        • Milanak ME
        • Moore T
        • Cobice D
        • Armour C.
        Sleep, a governor of morbidity in PTSD: a systematic review of biological markers in PTSD-related sleep disturbances.
        Nat Sci Sleep. 2020; 12: 545-562
        • Richards A
        • Kanady JC
        • Neylan TC.
        Sleep disturbance in PTSD and other anxiety-related disorders: an updated review of clinical features, physiological characteristics, and psychological and neurobiological mechanisms.
        Neuropsychopharmacology. 2020; 45: 55-73
        • Foster SN
        • Brock MS
        • Hansen S
        • et al.
        Sleep disorders related to deployment in active duty service members and veterans.
        Curr Pulmonol Rep. 2016; 5: 101-110
        • Elliott JE
        • Opel RA
        • Pleshakov D
        • et al.
        Posttraumatic stress disorder increases the odds of REM sleep behavior disorder and other parasomnias in Veterans with and without comorbid traumatic brain injury.
        Sleep. 2020; 43: zsz237
        • Good CH
        • Brager AJ
        • Capaldi VF
        • Mysliwiec V.
        Sleep in the United States military.
        Neuropsychopharmacology. 2020; 45: 176-191
        • Mysliwiec V
        • Gill J
        • Lee H
        • et al.
        Sleep disorders in US military personnel: a high rate of comorbid insomnia and obstructive sleep apnea.
        Chest. 2013; 144: 549-557
        • Moore BA
        • Tison LM
        • Palacios JG
        • Peterson AL
        Mysliwiec V. Incidence of insomnia and obstructive sleep apnea in active duty United States military service members.
        Sleep. 2021; 44: zsab024
        • McLay RN
        • Klam WP
        • Volkert SL.
        Insomnia is the most commonly reported symptom and predicts other symptoms of post-traumatic stress disorder in US service members returning from military deployments.
        Mil Med. 2010; 175: 759-762
        • Wright KM
        • Britt TW
        • Bliese PD
        • Adler AB
        • Picchioni D
        • Moore D.
        Insomnia as predictor versus outcome of PTSD and depression among Iraq combat veterans.
        J Clin Psychol. 2011; 67: 1240-1258
        • Caldwell JA
        • Knapik JJ
        • Shing TL
        • Kardouni JR
        • Lieberman HR.
        The association of insomnia and sleep apnea with deployment and combat exposure in the entire population of US army soldiers from 1997 to 2011: a retrospective cohort investigation.
        Sleep. 2019; 42: zsz112
        • Gehrman P
        • Seelig AD
        • Jacobson IG
        • et al.
        Predeployment sleep duration and insomnia symptoms as risk factors for new-onset mental health disorders following military deployment.
        Sleep. 2013; 36: 1009-1018
        • Wang HE
        • Campbell-Sills L
        • Kessler RC
        • et al.
        Pre-deployment insomnia is associated with post-deployment post-traumatic stress disorder and suicidal ideation in US Army soldiers.
        Sleep. 2019; 42: zsy229
        • Colvonen PJ
        • Almklov E
        • Tripp JC
        • Ulmer CS
        • Pittman JOE
        • Afari N.
        Prevalence rates and correlates of insomnia disorder in post-9/11 veterans enrolling in VA healthcare.
        Sleep. 2020; 43: zsaa119
        • Colvonen PJ
        • Masino T
        • Drummond SP
        • Myers US
        • Angkaw AC
        • Norman SB.
        Obstructive sleep apnea and posttraumatic stress disorder among OEF/OIF/OND veterans.
        J Clin Sleep Med. 2015; 11: 513-518
        • Rosen RC
        • Cikesh B
        • Fang S
        • et al.
        Posttraumatic stress disorder severity and insomnia-related sleep disturbances: longitudinal associations in a large, gender-balanced cohort of combat-exposed veterans.
        J Trauma Stress. 2019; 32: 936-945
        • Ryan MA
        • Smith TC
        • Smith B
        • et al.
        Millennium Cohort: enrollment begins a 21-year contribution to understanding the impact of military service.
        J Clin Epidemiol. 2007; 60: 181-191
        • Markwald RR
        • Carey FR
        • Kolaja CA
        • Jacobson IG
        • Cooper AD
        • Chinoy ED.
        Prevalence and predictors of insomnia and sleep medication use in a large tri-service US military sample.
        Sleep Health. 2021; 7: 675-682
        • Bastien CH
        • Vallieres A
        • Morin CM.
        Validation of the Insomnia Severity Index as an outcome measure for insomnia research.
        Sleep Med. 2001; 2: 297-307
        • Taylor DJ
        • Pruiksma KE
        • Hale WJ
        • et al.
        Prevalence, correlates, and predictors of insomnia in the US Army prior to Deployment.
        Sleep. 2016; 39: 1795-1806
        • Weathers FW
        • Litz BT
        • Herman DS
        • Huska JA
        • Keane TM.
        The PTSD Checklist (PCL): Reliability, validity, and diagnostic utility.
        in: Proceedings From the Annual Convention of the International Society for Traumatic Stress Studies, San Antonio, TX1993
        • American Psychiatry Association
        Diagnostic and Statistical Manual of Mental Disorders (DSM-IV).
        American Psychiatry Association, Washington, DC1994
        • Porter B
        • Hoge CW
        • Tobin LE
        • et al.
        Measuring aggregated and specific combat exposures: associations between combat exposure measures and posttraumatic stress disorder, depression, and alcohol-related problems.
        J Trauma Stress. 2018; 31: 296-306
        • National Research Council
        Returning Home From Iraq and Afghanistan: Assessment of Readjustment Needs of Veterans, Service Members, and Their Families.
        The National Academies Press, Washington, DC2013
        • Ware J
        • Kosinski M
        • Gandek B.
        SF-36 Health Survey: Manual and Interpretation Guide.
        QualityMetric Incorporated, Lincoln, RI2005
        • Kazis LE
        • Lee A
        • Spiro 3rd, A
        • et al.
        Measurement comparisons of the medical outcomes study and veterans SF-36 health survey.
        Health Care Financ Rev. 2004; 25: 43-58
        • Holmes TH
        • Rahe RH.
        The Social Readjustment Rating Scale.
        J Psychosom Res. 1967; 11: 213-218
        • Leggett A
        • Burgard S
        • Zivin K.
        The impact of sleep disturbance on the association between stressful life events and depressive symptoms.
        J Gerontol B Psychol Sci Soc Sci. 2016; 71: 118-128
        • Vandrey R
        • Babson KA
        • Herrmann ES
        • MO Bonn-Miller
        Interactions between disordered sleep, post-traumatic stress disorder, and substance use disorders.
        Int Rev Psychiatry. 2014; 26: 237-247
        • Jacobson IG
        • Ryan MA
        • Hooper TI
        • et al.
        Alcohol use and alcohol-related problems before and after military combat deployment.
        JAMA. 2008; 300: 663-675
        • Martindale SL
        • Morissette SB
        • Rowland JA
        • Dolan SL.
        Sleep quality affects cognitive functioning in returning combat veterans beyond combat exposure, PTSD, and mild TBI history.
        Neuropsychology. 2017; 31: 93-104
        • Kolaja CA
        • Porter B
        • Powell TM
        • Rull RP
        • Millennium Cohort Study T
        Multiple imputation validation study: addressing unmeasured survey data in a longitudinal design.
        BMC Med Res Methodol. 2021; 21: 5
        • Rubin DB.
        Multiple Imputation for Nonresponse in Surveys.
        Wiley, Inc., Hoboken, NJ2004
        • Littman AJ
        • Boyko EJ
        • Jacobson IG
        • et al.
        Assessing nonresponse bias at follow-up in a large prospective cohort of relatively young and mobile military service members.
        BMC Med Res Methodol. 2010; 10: 99
        • Mostafa T
        • Narayanan M
        • Pongiglione B
        • et al.
        Missing at random assumption made more plausible: evidence from the 1958 British birth cohort.
        J Clin Epidemiol. 2021; 136: 44-54
        • Armed Forces Health Surveillance Center
        Obstructive sleep apnea, active component, U.S. Armed Forces, January 2000-December 2009.
        Med Surveil Mon Rep (MSMR). 2010; 17: 8-11
        • Cicchetti DV
        • Feinstein AR.
        High agreement but low kappa: II. Resolving the paradoxes.
        J Clin Epidemiol. 1990; 43: 551-558
        • Landis JR
        • Koch GG.
        The measurement of observer agreement for categorical data.
        Biometrics. 1977; 33: 159-174
        • Cooper AD
        • Kolaja CA
        • Markwald RR
        • Jacobson IG
        • Chinoy ED.
        Longitudinal associations of military-related factors on self-reported sleep among U.S. service members.
        Sleep. 2021; 44: zsab168