Algorithm-Based Intermittent Hypoxia (SANA® Therapy) Reduces Headache and Improve Quality of Life in Patients with Persistent Post Concussive Symptoms (PPCS)

1. Introduction

Mild traumatic brain injury (mTBI), commonly referred to as concussion and commotio cerebri, is the most common neurological disorder worldwide, surpassing stroke, Alzheimer’s disease, and Parkinson’s disease in incidence [1]. While the prognosis for mTBI is generally favorable, a subset of patients develops persistent symptoms, often classified as persistent post-concussive symptoms (PPCS), post-concussion syndrome, or mild neurocognitive disorder (NCD) due to TBI [2,3,4,5]. A 2023 systematic review and meta-analysis estimated that the prevalence of PPCS after mTBI was 16% up to six months after initial presentation to emergency departments or trauma centers [6]. Additionally, a Swedish study reported that 12% of patients met the criteria for PPCS one-year post-injury [8]. Given that over half of the global population is projected to sustain at least one TBI in their lifetime [9,10], PPCS represents a significant public health challenge.

The symptomatic burden of PPCS can be substantial. The TRACK-TBI study reported a mean of nine symptoms at 12 months post-injury, as assessed by the Rivermead Post-Concussion Symptoms Questionnaire (RPQ) [11]. PPCS symptoms can be broadly categorized into physical (e.g. headaches, dizziness, nausea, sleep disturbances, fatigue, visual problems, and sensitivity to light and noise), cognitive (e.g. impaired memory, concentration, and slowed thinking), and mental (e.g. irritability, depression, anxiety, and frustration) [11,12]. The persistent symptomatic burden significantly reduces health-related quality of life (QoL) in both adults [13] and children [14], decreases life satisfaction [15], and impairs return to work [15,16,17,18], contributing to substantial personal and societal costs.

Development of persistent post-concussive symptoms (PPCS) has been proposed to be the result of biological, psychological, social, and environmental factors [1]. Of biological hypotheses axonal injury [19], neuro-inflammation [20], altered neurotransmission [21], oxidative stress [21,22], mitochondrial dysfunction [23], disruption of blood-brain barrier [24,25], and altered cerebral blood flow [26,27,28] have been proposed as underlying factors in different studies. These alterations might result in dysautonomia (especially during non-rest states) [29,30], pituitary dysfunction [31,32,33], impaired glymphatic drainage [34,35], neuronal loss [36], reduced brain functional connectivity [37], and altered neurometabolism [38]. Although the underlying mechanisms remain incompletely understood, the persistence of symptoms and high degree of treatment resistance following mTBI are likely driven by a heterogeneous and multifactorial pathophysiology.

Exposure to controlled sublethal levels of hypoxia, known as hypoxic conditioning, can increase resilience and promote recovery. In particular, interval-based hypoxia known as Intermittent Hypoxic Conditioning (IHC), which involves alternating periods of hypoxia (FiO₂ = 9-13%) with normoxia (FiO₂ = 21%) or hyperoxia (FiO₂ = 30-40%) (defined as Intermittent Hypoxic Conditioning (IHHC)), has demonstrated therapeutic potential across various pathologies [39,40,41]. Clinical trials have explored the effects of IHC in human diseases of the nervous system [40]-[47], cardiovascular system [48]-[62], respiratory system [63]-[67], and musculoskeletal system [44,46,47,68]. While no other studies to our knowledge have investigated the effects of this therapy on patients with PPCS, IHC have shown potential in reducing migraines [44] and dizziness [45], and improving quality of life [44,56,59,77], fatigue and exercise tolerance [46,58,63,64,66,77], blood flow [48,50,53,56]-[58,60]-[62], cognitive performance [42,43], and autonomic nervous regulation [62,65] in several other patient groups.

The underlying mechanisms have been extensively studied in cellular and animal models, with evidence suggesting that IHC/IHHC exerts its effects through reducing inflammation, restoring mitochondrial function, improving redox balance, modulating autonomic regulation, enhancing vascular function, and stimulating stem cell activity and tissue regeneration [39,69]. Given the overlap between the proposed pathophysiological mechanisms of PPCS and the biological effects of IHC/IHHC, we hypothesize that these therapies may offer a promising therapeutic approach for addressing the underlying causes driving persistent PPCS.

An appropriate hypoxic stimulus may be crucial for achieving beneficial treatment outcomes. Intermittent hypoxic-hyperoxic conditioning (IHHC) can be individualized using algorithm-based protocols to deliver the optimal hypoxic stimulus for each patient. The aim of the present study was to evaluate the effect of individualized IHHC on quality of life and headache-associated pain in patients suffering from long-term sequelae of mTBI.

2. Methods

2.1. Study Design

This study was conducted as an open-label cohort study at five medical clinics across Denmark. SANA® is a medical company specializing in individualized intermittent hypoxic/hyperoxic conditioning (IHHC), and all treatments were designed and administered by their medical staff. Prior to treatment initiation, all patients underwent a structured medical consultation with a physician, either in person or by telephone, including a comprehensive review of medical history.

Eligible patients were those experiencing persistent symptoms following one or more episodes of mild traumatic brain injury (mTBI), with the most recent injury sustained at least three months before treatment initiation. Minors (<18 years) could be included following individualized evaluation and with written consent from a parent or legal guardian. Exclusion criteria included pregnancy or a current diagnosis of malignant disease.

2.2. Data Collection and Questionnaires

Patients completed an online pre-treatment questionnaire 1 to 10 days before their first treatment session. Collected data included demographics (age, sex, height, weight), occupation, work absence (yes/no), and clinical details such as symptomatology, symptom duration, and prior treatments. Pain levels were assessed using the Numeric Rating Scale (NRS) for both rest and activity over the previous week (0 = no pain, 10 = worst imaginable pain). Additionally, patients completed the Short Form-36 (SF-36) health-related quality-of-life questionnaire. The diagnosis of PPCS was determined by the physician based on medical history and symptomatology.

2.3. Treatment Protocol

The SANA® Therapy sessions were administered with patients seated in a reclining chair and breathing through a tightly fitted oxygen mask. The IHHC protocol was personalized via an algorithm (version 1.0) developed from data on 756 patients previously treated under physician supervision. The algorithm-tailored session parameters included the duration of intervals, the total hypoxic exposure, and the target oxygen saturation for each session, based on the patient-specific factors as symptom severity, baseline pain levels (NRS), and SF-36 scores.

Each session comprised hypoxic intervals (3–8 minutes) followed by hyperoxic intervals (1–3 minutes). Inspiratory oxygen concentrations were maintained at 8.5–13% during hypoxia and 34–36% during hyperoxia. A pulse oximeter continuously monitored oxygen saturation, targeting levels between 77% and 84% determined by the algorithm. Session durations ranged from 18 to 42 minutes, with the total number of sessions determined by symptom progression. The interval between sessions varied according to algorithmic recommendations, typically starting with an interval of 7 days between the first three sessions, then extending to 10 to 21 days between subsequent treatments. Treatments were administered using certified medical hypoxicators (Hypoxbreath®, INVATIO, Hannover, Germany and CellOxy, TUR, Rostock, Germany).

2.4. Patient Inclusion and Follow-Up

Data were sourced from the SANA® database, comprising all patients treated at SANA® Clinics in Denmark between January 1, 2020, and October 31, 2025, with a minimum follow-up period of six-months. Patients were either self-referred or referred by insurance companies for rehabilitation. Inclusion required a minimum symptom duration of three months. Patients who completed less than three sessions or whose symptoms lasted less than three months were excluded from the statistical analysis.

Follow-up questionnaires were sent and completed online at six weeks and six months post-treatment initiation. Each patient received a single questionnaire per time point, and no reminder emails were sent. Responses at these intervals were analyzed to assess treatment efficacy, and patients who completed the respective questionnaires were included in the statistical analysis for those time points.

2.5. Statistical Analysis

Data were analyzed using Prism 10 (Version 10.6.0, GraphPad Software LLC). Normal distribution was assessed via QQ-plots of residuals. Descriptive statistics summarized patient demographics, symptom duration, and baseline measures. Group comparisons of demographic variables were conducted using one-way ANOVA with Sidak’s multiple comparisons correction where applicable. The Chi-square test was used for categorical demographic comparisons (e.g., sex distribution).

Changes in health-related quality of life (SF-36 scores) were evaluated using paired t-tests for within-cohort change (baseline to follow-up) with Sidak’s correction across the eight SF-36 domains. As a HRQoL summary, Physical Component Summary (PCS) and Mental Component Summary (MCS) scores were calculated using the Ware 1994 algorithm with US 1990 norm-based factor weights [70], which has been shown to be highly equivalent to country-specific algorithms in Danish data[71]. As a secondary summary, the SF-36 total score was computed as the sum of the eight domains (health transition (HT) excluded, as HT is a single transition item rather than a domain). Between-cohort baseline comparisons used Welch’s two-sample t-test. Pain scores (NRS, 0–10) were analyzed for patients reporting an initial pain level of ≥3 at rest or during activity. One-way ANOVA with Sidak’s correction was applied for multiple comparisons within each cohort.

A significance threshold of P < 0.05 was applied to all statistical analyses. Data were reported as means with 95% confidence intervals (CI), expressed both as absolute values and percentages relative to baseline, to assess the precision of the estimates.

2.5. Ethics

The study was approved by the local ethics committee under the Danish National Committee of Research Ethics (#1-10-72-274-21). The study was conducted in accordance with the Declaration of Helsinki, and all participant data were handled in compliance with the EU General Data Protection Regulation (GDPR).

3. Results

3.1. Patients and Treatment

A total of 553 patients with persistent post-concussive symptoms (PPCS) were identified in the SANA® consecutive database in the inclusion period. 19 patients were excluded due to symptom duration of less than three months, and three patients were excluded due to receiving less than three treatments (Figure 1). The first criterion was included to investigate the effects of IHC on chronic sequalae of mTBIs. In the remaining cohort, 158 patients completed the 6-week follow-up questionnaire, and were included in this study, constituting a 6-week follow-up cohort, with a mean age of 40.8 years (range: 11–79 years). The female-to-male ratio was 70:28 (with 2 patients of unspecified sex), and the mean symptom duration was 28.5 months. Sick leave from work or school at baseline was reported by 49.4% of patients. The 6-month follow-up questionnaire was completed by 44 patients, yielding a response rate of 27.8%. This formed a 6-week follow-up cohort (n=158) and a 6-month follow-up cohort (n=44) for subsequent paired analyses (Figure 1).

Group demographics are presented in Table 1. There was no significant difference between groups in age, sex, anthropometrics, symptom duration, or number of prior treatments. Sick leave at baseline were the only measured parameter where there was significant difference between groups (49.4% vs. 31.8%). Patients received a mean of 9.07 treatment sessions in the 6-week follow-up cohort and 9.86 in the 6-month follow-up cohort (median 7, range 3–35); 61.4% received between 4 and 8 sessions.

Prior to SANA® therapy, 98.7% of patients in the 6-week follow-up cohort reported that they had tried other treatment modalities with an average of 5.23 treatment types (median 5, range 0-20). The most common attempted treatment was physiotherapy (75.9%) and non-prescription analgesics (69.0%). A list of the most common prior treatments is provided in Table 2.

The most frequently noted side effects were mild headaches and tiredness that persisted for 12–24 hours. These symptoms generally subsided after the first few sessions as patients became accustomed to the hypoxic exposure. No serious adverse reactions were observed.

3.2. Pain

In the 6-week follow-up cohort, 86.1% of patients reported any pain with an NRS-score of above 0 at baseline. Subsequent pain analysis included patients with a baseline Numeric Rating Scale (NRS) score of ≥3 at rest and/or during activity. At baseline, 109 patients (69.0%) in the 6-week follow-up cohort reported pain with NRS ≥3 at rest with a mean intensity of 5.41 [95% CI, 5.09; 5.74]. During activity, 122 patients (77.2%) reported pain with NRS ≥3 with a mean intensity of 6.05 [95% CI, 5.70; 6.39]. Pain levels were significantly higher during activity compared to rest at baseline in both the 6-week and the 6-month cohorts (p < 0.01). In the 6-week follow-up cohort at baseline, 81.7% reported headache as a pain source, 50.0% neck pain, and 6.7% shoulder pain. Regardless of NRS-score, 80.4% of patients reported headache as a post-concussive symptom.

From baseline to the 6-week follow-up, pain at rest decreased by 1.30 NRS-points [95% CI, 0.90; 1.71] (mean pain reduction of 24.0%; p < 0.001), and pain during activity decreased by 1.27 points [95% CI, 0.88; 1.66] (mean pain reduction of 21.0%; p < 0.001). At the 6-month follow-up, pain was reduced at rest by 2.30 NRS-points [95% CI, 1.50; 3.11] (mean pain reduction of 44.4%; p < 0.001; n = 33) and during activity by 2.57 points [95% CI, 1.69; 3.45] (mean pain reduction of 41.3%; p < 0.001; n = 35) (Figure 2). Pain reduction was significantly greater at the 6-month than at the 6-week follow-up at rest (Δ = –1.00 NRS-points; p = 0.035) and during activity (Δ = –1.30 NRS-points; p = 0.009).

3.3. Quality of Life

Quality of life was evaluated using the SF-36 health survey (Table 4), which consists of 36 questions that measure 8 health domains: physical functioning, role-physical, bodily pain, general health, vitality, social functioning, role-emotional, and mental health. An additional measure, self-reported health transition over the past year, was collected but not included in the total score. Scores for each domain range from 0 to 100, with higher scores indicating better health [56].

At baseline, values were generally low, with average total scores of 384 in the 6-week follow-up cohort and 394 in the 6-month follow-up cohort. For comparison, the total SF-36 score for the Danish background population has been estimated to be 645 [74]. The lowest baseline scores were observed in the domains role-physical (scores of 14.2 and 12.5 in the 6-week and 6-month cohort, respectively) and vitality (scores of 28.1 and 28.3, respectively). There were no significant differences between the 6-week and 6-month follow-up cohorts at baseline in any SF-36 domain (Welch’s two-sample t-test, all p > 0.05).

Significant improvements were observed in all domains from baseline to the 6-week follow-up. In the 6-month follow-up group, significant improvements were seen in all domains except for the domain general health (Figure 3). Using the total score as a surrogate for overall quality of life, improvements from baseline of 90.5 points (23.6%) (p=5.6x10⁻¹⁷) at 6 weeks post-treatment initiation and 131.7 points (33.4%) (p=9.0×10⁻⁷) at 6 months post-treatment initiation were observed.

To assess the patients’ self-perceived evolution in health status, the domain health transition was included. Health transition scores below 50 indicate a perceived decline in health compared to one year ago, a score of 50 signifies no change, and a score above 50 reflects an improvement in health over the past year. Health transition scores improved significantly from a score below 50 at baseline to above 50 at the 6-week and 6-month follow-ups (p < 0.001). The mean health transition score increased from 42.2 at baseline to 58.2 at the 6-week follow-up, with a mean increase of 16.0 points [11.9; 20.1] (p < 0.001). In the 6-month follow up cohort, the health transition score increased with 19.3 points [11.8;26.9] (p < 0.001) from 47.2 to 66.5.

3.4. Physical and Mental Component Summaries

To complement the per-domain SF-36 analysis, Physical Component Summary (PCS) and Mental Component Summary (MCS) scores were computed. PCS and MCS are weighted, summarized scores of the eight SF-36 domains and are interpreted relative to U.S. population norms of 50, with lower scores reflecting greater impairment [70]. The Danish population norms have been calculated to 51.8 and 56.1 for PCS and MCS, respectively [72]. A change of approximately 2–5 points in SF-36 PCS or MCS scores is generally considered clinically meaningful [75,76].

In the 6-week follow-up cohort, PCS improved from 39.05 ± 8.63 at baseline to 41.4 ± 8.81 at 6 weeks (mean change +2.39, 95% CI 1.23;3.55, p = 8.7×10⁻⁵). MCS improved from 38.6 ± 11.0 to 45.3 ± 10.5 (mean change +6.74, 95% CI 5.09;8.39, p = 2.6×10⁻¹³). In the 6-month follow-up cohort, PCS improved from 37.9 ± 7.83 to 43.1 ± 9.06 (mean change +5.26, 95% CI 3.27 to 7.25; p = 5.5×10⁻⁶). MCS improved from 41.0 ± 10.2 to 48.8 ± 10.1 (mean change +7.78, 95% CI 4.22 to 11.34; p = 1.0×10⁻⁴). At baseline, the 6-week cohort lay approximately 1.6 SD below the Danish PCS norm and 1.7 SD below the MCS norm, consistent with the reduced physical and mental health burden expected in this patient group.

4. Discussion

Each year, fifty-five million individuals are estimated to sustain a mild traumatic brain injury globally [10]. mTBIs have often been referred to as the “silent epidemic”, and while it has the highest incidence of neurological diseases worldwide [1], much remains unknown in the pathophysiological understanding and in the treatment of the disease and its sequelae. This study presents a potential treatment option for patients suffering from long-term sequelae after mild traumatic brain injuries. This current study investigated the effects of algorithm-based, personalized intermittent hypoxia-hyperoxia conditioning (IHHC) in a cohort of 158 PPCS-patients with a mean symptom duration of 28.5 months. At baseline, the cohort had a 40.5% decreased health-related quality-of-life score measured by SF-36, when compared to the general Danish population [74]. Six weeks and six months after a treatment initiation with IHHC, health-related quality of life was significantly increased by 23.6% and 33.4%, respectively. Notably, the SF-36 mental component (MCS) improved from a level corresponding to moderate-to-severe impairment at baseline to only mildly reduced function at the 6-month follow-up, a clinically meaningful shift that reflects a substantial change relative to the general population norm. At baseline, 69.0% of patient experienced pain with an NRS-score ≥3 either at rest or during activity with headache as the largely predominant source of pain. After the treatment regiment, pain intensity was reduced significantly by 19.5-44.4%.

Post-traumatic headache (PTH) is a common sequela of mTBI. In the current cohort, 81.7% of patients reported PTH with an intensity of NRS ≥3, which is higher than prior studies estimating a headache prevalence of 34-58% one year post-trauma [11,12,77]. This might reflect SANA®’s primary focus on pain patients. PTH most frequently presents as a migraine-like headache, followed by a tension-type phenotype [78,79], with evidence suggesting shared pathophysiological mechanisms between migraine and post-traumatic migraine-like headache [80,81]. The pathophysiology of PTH may involve a) impaired descending pain modulation, b) neurometabolic changes, c) cortical spreading depression, d) calcitonin gene-related peptide (CGRP)-dependent mechanisms, and e) neuroinflammation [81]. Imaging studies have revealed altered cerebral blood flow (CBF) following mTBIs, correlating with post-concussive symptoms [1,82,83,84]. These CBF changes may drive the functional remodeling, exacerbate the metabolic crises, and activate the trigeminal sensory system [81] leading to headache and other PPCS symptoms.

Early treatment may be crucial in preventing the chronification of symptoms. Persistent post-traumatic headache (lasting more than 3 months) is largely attributed to central sensitization, which arises from continuous transmission from first-order neurons to second- and third-order neurons within the central nervous system [85]. Interventions that reduce initial pain transmission may curb central sensitization. Notably, targeted therapies against CGRP signaling have been shown to prevent allodynia in concussed rodents [86]. CGRP is also believed to be central in the signaling of nausea, photo- and phonophobia [87,88,89]. A 2020 randomized controlled trial found that an 8-week IHC regimen reduced migraine frequency, pain, depression, and anxiety while improving quality of life in migraine patients [44]. This was associated with decreased CBF velocity in the middle cerebral and basilar arteries accompanied by lower VEGF and CGRP levels [44]. Together it can be hypothesized, that not only may IHC be a potential treatment of PPCS, but the use of IHC may also be beneficial in the early phases of traumatic brain injury (TBI) to mitigate persistent symptoms by reducing CGRP activity. In this current study, the mean symptom duration was 28.5 months, reflecting the chronic and persistent nature of symptoms in the investigated cohort. Future research should explore the potential of IHC in the early stages of mTBIs to prevent symptom chronification.

This study has several limitations. First, the findings rely exclusively on self-reported questionnaires, which carry inherent biases. However, in the context of long-term sequelae after mTBI – where validated objective biomarkers are lacking – patient-reported outcomes remain among the most relevant indicators of treatment efficacy. Another limitation is the lack of use of headache-specific questionnaires to quantify headache frequency, intensity, or migraine attacks. Future studies on IHC’s effects on post-concussive symptoms should therefore include headache-specific questionnaires as well as the Rivermead Post-Concussion Symptoms Questionnaire, as PPCS encompasses a broad range of symptoms beyond headache.

Second, the retrospective design precluded blinding and the inclusion of a placebo-controlled group, limiting causal inference and raising the possibility that the observed improvements may reflect placebo effects. However, it is noteworthy that the ‘health transition’ scores of the SF-36 evaluation in both cohorts were at or below 50, indicating respectively health stagnation or decline over the preceding year. Participants had a mean symptom duration of 28.5-33.2 months and had previously attempted an average of 5.16-5.23 treatment modalities before initiating SANA® Therapy, underscoring the chronic and treatment-resistant nature of symptoms in this cohort. All patients were either self-paying or referred by an insurance company with also may have attributed to placebo responses due to self-justification and expectations.

Finally, the low follow-up rates in the 6-week and 6-month cohorts could have been increased by sending reminders to participants, and it remains unknown if the respondents at follow-up disproportionately represented patients who improved or those who did not.

One of the main strengths of this study is its heterogeneous cohort, which enhances the external validity of the findings. By applying broad inclusion criteria, the study population reflects the diverse and multifaceted symptomatology of patients with PPCS, encompassing psychological, physical, and social dimensions. This real-world representativeness increases the generalizability of the results to the broader population of patients with persistent post-concussive symptoms. The symptom burden in PPCS likely arise from a complex interplay of genetic, physical, psychological, and social factors [1]. Consequently, complete resolution of PPCS may require individualized assessment and interdisciplinary treatment approaches [90] combining relevant elements such as physiotherapy [91], cognitive behavioral therapy and psychoeducation [92,93], reposition maneuvers for benign paroxysmal positional vertigo [94], aerobic training [91,95], or pharmacological treatment [80,96]. Future studies should investigate whether combining such modalities with IHHC provides greater benefit than IHHC alone. Both efficacy and cost-effectiveness of these treatments must be evaluated in order to resolve PPCS for the most patients possible.

5. Conclusions

In conclusion, this study demonstrates the potential of individualized, algorithm-based IHHC-treatment (SANA® Therapy) in reducing headache-associated pain and improving quality of life in this treatment-resistant PPCS-cohort with a symptom duration of 26.8 to 33.2 months of a stagnant or worsening nature. The findings indicate IHHC’s potential as a clinically relevant and cost-effective treatment-option for PPCS patients.

6. Patents

No patents are disclosed in this manuscript. However, the SANA® Therapy algorithms are proprietary, confidential, and remain the exclusive intellectual property of SANA Medical Systems ApS, Risskov, Denmark.