The Effect of a Low-Glycemic Index Diet on Postprandial Hypotension in Individuals With Chronic Spinal Cord Injury: Results From a Pilot Study

Abstract

Background:

One in two individuals with spinal cord injury (SCI) experiences postprandial hypotension (PPH), a decline (>20 mm Hg) in systolic blood pressure (SBP) within 2 hours after eating. Consuming meals with a low glycemic index (GI) could prevent or lessen PPH.

Objectives:

To determine the effect of a low-GI diet on PPH and postprandial glucose and insulin in individuals with chronic SCI (>1 year postinjury).

Methods:

Eleven participants (6 males, 5 females; age 43 ± 11 years) with chronic SCI (C4-C7, 7; T4-T12, 4) took part in a randomized crossover study (low GI vs. high GI). On each occasion, BP, glucose, and insulin were measured in the fasted state and for 2 hours after consuming a breakfast meal (60% carbohydrate, 28% fat, 12% protein) in laboratory-controlled conditions. Participants wore an ambulatory BP monitor and continuous glucose monitor for 3 days at home, and consumed study meals that were macronutrient-matched across conditions.

Results:

The maximum decrease in systolic blood pressure (SBP) following the laboratory-controlled breakfast meals tended to be lower in the low-GI (14 ± 12 mm Hg) compared to the high-GI (24 ± 25 mm Hg) diet (d = 0.52, P = .056). Serum glucose (P < .01) and insulin (P = .026) concentrations were lower at 30 minutes in the low-GI diet. In the home setting, peak glucose concentrations were lower after lunch (P = .011) and dinner (P < .01) in the low-GI diet.

Conclusion:

A low-GI meal may be an effective solution to reduce the magnitude of PPH and peak glucose concentrations in individuals with chronic SCI.

Introduction

There are an estimated 20.6 million individuals living with a spinal cord injury (SCI) worldwide.¹ Individuals surviving the first year of injury have a substantially reduced life expectancy than the general population.² Cardiovascular disease (CVD) is the second leading cause of death (after respiratory diseases) for this population.^3,4 Moreover, compared to the general population, the incidence of CVD is up to three times higher in individuals living with SCI.^5,6 In addition to the traditional CVD risk factors, SCI-specific mechanisms contributing to the increased prevalence of CVD include autonomic dysfunction, chronic inflammation, physical deconditioning, sarcopenia, and altered metabolic homeostasis.⁷

Autonomic dysfunction in individuals with SCI also appears to increase the risk of postprandial hypotension (PPH). PPH is defined as a decrease of 20 mm Hg in systolic blood pressure (SBP) within 2 hours of meal ingestion or <90 mm Hg after having been >100 mm Hg before the meal.⁸ In non-SCI populations, PPH has been associated with a twofold higher CVD mortality rate, with the magnitude of the drop in SBP linearly associated with all-cause mortality rate.^9,10 The exact mechanism by which PPH increases CVD risk is unclear, but it is possible that these hypotensive events lead to myocardial ischemia.¹¹ PPH occurs in approximately 50% of individuals with SCI, with a greater occurrence in older individuals, those with a higher level of injury, and those with complete SCI.¹² Unfortunately, most cases are asymptomatic, making diagnosis difficult.

The pathogenesis of PPH is linked to autonomic dysfunction. Specifically, in response to eating, blood pools in the small intestines to facilitate digestion and nutrient absorption, leading to a reduction in circulating blood volume and a drop in BP.¹³ In healthy individuals without SCI, this hypotensive effect of eating is buffered by the activation of the autonomic nervous system, leading to an increased heart rate, cardiac output, and peripheral vascular resistance.¹⁴ However, in individuals with SCI, autonomic dysfunction and impairment of the sympathetic gastrovascular reflex may blunt the gastric distension-elicited vasoconstriction response to eating, leading to PPH.^12,15 Glycemic disorders in individuals with SCI are likely compounding this problem. The prevalence of impaired glucose intolerance¹⁶ in individuals with chronic SCI is considerably higher than in the general population. This increases the risk of CVD¹⁷ and likely exacerbates PPH due to higher insulin secretion, a known vasodilator.¹⁸

Lowering the glycemic index (GI) of meals could be an effective solution to PPH and impaired postprandial glycemia. GI is a value from 0 to 100 that reflects how much a specific food raises blood glucose compared to a food with an arbitrarily set GI of 100 (e.g., white bread or glucose). It is possible to manipulate the GI of meals to reduce postprandial blood glucose and insulin responses.^19,20 If these responses can be ameliorated, and the rate of glucose entering the small intestine reduced, the magnitude of PPH may be reduced, thereby reducing CVD risk. Therefore, the aim of this study was to determine the effect of a low-GI diet on PPH and postprandial glycemia in individuals with chronic SCI. The hypothesis was that compared to a high-GI (control) diet, consuming a low-GI diet will reduce the magnitude of PPH and postprandial glycemia. The primary outcome measure was the maximum drop in SBP under laboratory-controlled conditions between the high-GI and low-GI meals.

Methods

This study was approved by the Institutional Review Board at The Ohio State University (2023H0151) on July 19, 2023. All participants provided informed consent and were monetarily compensated. The study was registered as a clinical trial at ClinicalTrials.gov (https://clinicaltrials.gov) under the identifier NCT05334342.

Participants

An a priori sample size calculation (β = 0.80, α = 0.05) determined a recruitment target of 12 participants to detect an effect size of Cohen d = 1.0 for the magnitude of PPH between the high- and low-GI conditions under laboratory-controlled conditions. This effect size was calculated from an 18.0 ± 18.9 mm Hg difference (vs. placebo) in PPH following acarbose administration (an α-glucosidase inhibitor that decreases glucose absorption in the small intestine) in patients with autonomic failure.²¹

A total of 11 individuals completed all study visits. Participant characteristics are presented in Table 1. Participants were eligible if they met all the following criteria: age between 18 and 65 years old, chronic (>1 year postinjury) spinal cord lesion at or below the fourth cervical level (C4), self-reported use of a wheelchair for activities of daily life, and not pregnant or breastfeeding. Individuals who self-reported the use of insulin to treat diabetes and a neurological condition other than SCI were excluded.

Table 1.

Participant characteristics (N = 11)

Age, years	43 ± 11
Sex
Male	6
Female	5
AIS classification
A	4
B	5
C	2
LOI
Cervical	7
Thoracic	4
TSI, years	17 ± 13
BMI, kg/m2	31 ± 6

Note: AIS = American Spinal Injury Association Impairment Scale; BMI = body mass index; LOI = level of injury; TSI = time since injury.

Study design

Participants visited the Clinical Research Center at The Ohio State University on three separate occasions (one pre-experimental visit and two experimental conditions) in a single-blinded randomized crossover design. Before all three visits, participants were instructed to refrain from strenuous exercise, caffeine, and alcohol the day prior and arrive after an overnight fast (10-12 hours). All visits were separated by ≥7 days to reduce the risk of a carryover effect. The study design is shown in Figure 1.

Figure 1.

Schematic of study design and procedures. BP = blood pressure; CGM = continuous glucose monitoring; GI = glycemic index; PARA-SCI = Physical Activity Recall Assessment for People with Spinal Cord Injury; PPH = postprandial hypotension; RMR = resting metabolic rate.

Pre-experimental visit

Body mass was measured using wheelchair scales (Scale-Tronix 6722, Welch Allyn, NY), with the wheelchair weighed separately, and height was self-reported. Body mass index (BMI) was calculated as body mass (kg) divided by height (m) squared. Participants transferred to a medical bed and were fitted with an automated sphygmomanometer (OnTrack 90227, Spacelabs Healthcare, Snoqualmie, WA) to measure brachial blood pressure (BP) and heart rate. Measurements began once the participant was in a comfortable supine position and were obtained every 6 minutes, for at least four times (i.e., at least 18 minutes) in the fasted state. During this time, resting metabolic rate (RMR) was determined using indirect calorimetry (Q-NRG, COSMED, Rome, Italy). This was performed using the canopy system, and participants rested quietly for 20 minutes during this measurement. RMR was calculated as the mean of the lowest 5-minute moving average period.

Participants then consumed a standardized drink (Boost^®, Nestlé; Vevey, Switzerland), which contained a similar macronutrient content (78.3 g carbohydrate, 2.8 g fat, 20.6 g protein) and energy content (414 kcal) to tolerance meals previously used for the diagnosis of PPH.⁸ BP and heart rate were measured every 6 minutes after the start of drinking for 2 hours. Due to the effect of sitting upright on BP in individuals with SCI,²² positional changes were carefully controlled. Specifically, participants laid supine during the fasting BP measurements, sat upright while drinking, and then laid supine for the remainder of the 2-hour postprandial period. Measurements taken while drinking and/or sitting up were excluded from the analysis.

At the end of this visit, participants were interviewed using the Physical Activity Recall Assessment for People with Spinal Cord Injury (PARA-SCI)²³ questionnaire to calculate physical activity energy expenditure.²⁴ Physical activity energy expenditure was added to RMR to calculate the total daily energy requirements for each participant.

Experimental conditions

Participants completed two experimental conditions (low GI and high GI) in a randomized order. A randomization sequence for diet assignment order was generated prior to recruitment using a random permuted blocks strategy. Upon arrival, participants were transferred to a medical bed, and an intravenous catheter was inserted into an antecubital vein. Participants were fitted with the same BP machine as the pre-experimental visit, with at least four measurements taken at the same serial intervals (i.e., every 6 minutes). Participants then consumed a breakfast meal corresponding to the trial arm: low GI (GI = 55) or high GI (GI = 95). The meal was matched for energy content and macronutrient composition (Table 2), and participants were given 15 minutes to finish eating (white bread as reference).

Table 2.

Energy and macronutrient breakdown of meals for low- and high-glycemic index diets

	% of Total energy requirements	Carbohydrate (%)	Fat (%)	Protein (%)
Breakfast	20	60	28	12
Lunch	30	54	30	16
Dinner	40	50	32	18
Snack	10	68	26	6

Blood samples were collected in the fasted state and 10, 30, 60, 90, and 120 minutes after the start of eating. Samples were left to clot at room temperature for 30 minutes, centrifuged (1932 G for 10 minutes) before serum was aliquoted, and then stored at -80 °C until analysis. Serum glucose concentrations were determined using an automated analyzer (Beckman Coulter AU5800, Beckman Coulter Inc., Brea, CA). Serum insulin concentrations were determined using electrochemiluminescence assays (Meso Scale Discovery, Rockville, MD).

Upon leaving the laboratory, participants were provided with food for the at-home phase of the study (i.e., the remainder of day 1 and days 2 and 3). The macronutrient breakdown and menu for each diet is shown in Tables 2 and 3, respectively. Participants were instructed to separate each meal by a minimum of 2 hours. Participants were provided with a menu checklist to record the time they began consuming each meal and were given scales to weigh and record any leftovers. Compliance was assessed using the total grams remaining per meal.

Table 3.

Menus for high-glycemic index (GI) and low-GI diets

High GI (GI=95)	Low GI (GI=55)
Breakfast - Cereal	Breakfast - Cereal
Corn flakes	Bran flakes
2% milk	Almond milk
White bread	Almonds
Butter	Canned peaches
Raisins
Lunch - Chicken salad sandwich	Lunch - Mediterranean salad
White bread	Lettuce, mixed greens
Chicken breast	Lentils, chickpeas, bulgar mix
Mayonnaise	Tomato, sun-dried
Grapes	Feta cheese
Onion	Bacon
Salt and pepper	Avocado
Spinach	Hummus
Feta cheese	Olive oil
Craisins	Apple
Olive oil and balsamic vinegar
Dinner – Teriyaki chicken	Dinner - Southwest grain bowl
Chicken breast	Barley (pearled)
Teriyaki marinade	Beans (black and pinto)
Broccoli	80% lean ground beef
Sesame seeds	Taco seasoning
Canola oil	Peppers (green and red)
White rice	Canola oil
Gatorade powder	Onion
	Lime juice
	Tomato, raw
	Butter
	Sugar, white granulated
Snacks	Snacks
KIND Healthy Grains bar	Dried apples
Gatorade powder	Peanut butter

During the at-home phases, the ambulatory BP device (OnTrack 90227) was programmed to measure BP every 15 minutes during the daytime and every 60 minutes during sleep. This schedule was individualized for each participant but was kept the same for both conditions. Participants were instructed to minimize movement when measurements were being obtained and to ensure they began eating dinner at least 2 hours before sleeping. The ambulatory device used has been validated for use in clinical settings²⁵ and is programmed to retake BP 3 minutes after an unsuccessful scheduled reading. Ambulatory BP data was considered valid and included for analysis if two or more successful readings were recorded in 60 minutes before meal consumption and eight successful post-meal readings were recorded in the 2 hours after meal consumption. Participants were fitted with a continuous glucose monitor (CGM; Dexcom G6 Pro, Dexcom Inc., San Diego, CA) on their abdomen to measure glucose in the at-home setting before leaving the laboratory. Measurements of interstitial glucose concentration were obtained every 5 minutes, with participants blinded to their data. Diet logs were used to identify the 2-hour postprandial period for BP and CGM data. One participant did not complete both at-home conditions, and one participant's meal log was not returned, and therefore nine individuals are included in the at-home analysis.

Statistical analysis

Incremental area under the curve (iAUC) was calculated for glucose (in-lab and at-home) and insulin (in-lab only) for the 2 hours after the start of each meal. One-tail paired t tests were used to determine if significant differences were present between conditions for the magnitude of PPH and glucose and insulin iAUC. Linear mixed-model analyses were constructed to (1) test the effect of diet, time, and their interaction, randomization sequence, with the participant as a random effect, for the in-lab glucose and insulin responses over the 2-hour period and (2) test the effect of diet, day, meal, diet and meal interaction, randomization sequence, and study meal compliance with the participant as a random effect. Pairwise post hoc comparisons were performed using the Tukey-Kramer multiple comparisons method to estimate diet differences. Statistical model assumptions were validated before data analysis. Statistical tests were two-sided (unless otherwise noted), and P ≤ .05 was considered statistically significant. Statistical analyses were performed using the SAS version 9.4 (SAS Institute, Inc.). Data presented are mean ± standard deviation unless otherwise stated. Effect sizes (Cohen d) were calculated for selected variables and can be interpreted as follows: trivial effect 0.0-0.19, small effect 0.20-0.49, medium effect 0.50–0.79, and large effect ≥0.80. The study was conducted as a pilot study, and therefore the results should be considered exploratory rather than confirmatory.

Results

Seven of the eleven participants recruited to the study exhibited PPH at the screening visit. For these participants, the maximum reduction in SBP was 33 ± 14 mm Hg. The average total daily energy intake requirements for all participants were 1970 ± 589 kcal.

Laboratory-controlled breakfast (day 1)

Participants took longer to consume the lowGI breakfast (13 ± 3 min) compared to the high-GI breakfast (10 ± 3 min; P < .01). The maximum decrease in SBP following the laboratory-controlled breakfast meals tended to be lower following the lowGI breakfast compared to the high-GI breakfast (d = 0.52, P = .057; Figure 2). Two participants experienced PPH during the low-GI diet, and three participants experienced PPH during the high-GI diet.

Figure 2.

Maximum decrease in systolic blood pressure (SBP) following the high-glycemic index (GI) and low-GI breakfasts during controlled laboratory settings (n = 11). Data presented as mean and 95% CI (upper limit). Individual responses noted.

When fasting concentrations of serum glucose and insulin were averaged across both experimental conditions, one participant could be classified as having diabetes (fasting glucose ≥126 mg/dL), two participants could be classified as having prediabetes (fasting glucose ≥100 mg/dL), and eight participants could be classified as having significant insulin resistance (Homeostatic Model Assessment of Insulin Resistance [HOMA-IR] ≥2.5).^26,27

There was a significant diet and time interaction effect for serum glucose (P = .033; Figure 3) and insulin (P = .039; Figure 4). Bonferroni post hoc analysis revealed that at 30-minute post-meal consumption, glucose (P < .01; Figure 3) and insulin (P = .026; Figure 4) were higher in the high-GI compared to low-GI diet. Peak insulin (d = 0.46, P = .061) and glucose (d = 0.40, P = .062) concentrations tended to be lower in the low-GI compared to high-GI diet. There was no difference in glucose iAUC between the low-GI (1923 ± 1101 mg/dL per 120 min) and high-GI (2281 ± 1114 mg/dL per 120 min) breakfasts (d = 0.32, P = .144). There was no difference in insulin iAUC between the low-GI (5105 ± 3547 μIU/mL per 120 min) and high-GI (6398 ± 2731 μIU/mL per 120 min) breakfasts (d = 0.29, P = .079).

Figure 3.

Glucose concentrations in the 2 hours following the high-glycemic index (GI) and low-GI breakfasts during controlled laboratory settings (n = 11). Data presented as mean ± 95% CI. *P ≤ .05.

Figure 4.

Insulin concentrations in the 2 hours following the high-glycemic index (GI) and low-GI breakfasts during controlled laboratory settings (n = 11). Data presented as mean ± 95% CI. *P ≤ .05.

At-home setting

Individuals consumed a lower proportion of the total weight of food provided in the low-GI (81% ± 19%) compared to the high-GI diet (90% ± 15%) (n = 9, P = .018).

Only 83 of 144 meals (57.6%) had valid ambulatory BP data. This included 43 from the high-GI diet and 40 from the low-GI diet. Of the 83 valid meal recordings, there were 30 meals that met the criteria for PPH, including 17 during the high-GI diet and 13 during the low-GI diet, for eight of the nine participants. There was no main effect of diet (P = .994) or interaction effect (Diet x Meal) (P = .653) for the maximum decrease in SBP in the 2 hours after meal consumption (Figure 5).

Figure 5.

Maximum decrease in systolic blood pressure (SBP) following the high-glycemic index (GI) and low-GI meals during the at-home setting (n = 9). Data presented as mean ± 95% CI.

There was a significant interaction effect for peak glucose concentrations from the CGM (n = 9, P = .050), with post hoc comparisons showing significantly higher peak concentrations for the high-GI diet compared to the low-GI diet for lunch (P = .011) and dinner (P < .01) but not for breakfast (P = .716) (Table 4). There was no main effect of diet (P = .152) or interaction effect (Diet x Meal) (P = .297) for glucose iAUC during the at-home phase.

Table 4.

Peak glucose and incremental area under the curve (iAUC) during high-glycemic index (GI) and low-GI diets during at-home phase (n = 9)

	High GI	Low GI	P value
Peak glucose, mg/dL
Breakfast	164 (143-185)	168 (148-187)	.716
Lunch	141 (122-160)	123 (104-142)	.011
Dinner	166 (146-185)	143 (124-162)	<.01
iAUC, mg/L per 120 min
Breakfast	39.8 (29.7-49.9)	41.9 (32.9-60.0)	.678
Lunch	34.3 (25.2-43.4)	26.7 (18.0-35.4)	.073
Dinner	41.7 (32.7-50.8)	35.6 (26.9-44.3)	.152

Note: Values provided as mean (95% CI). Bold indicates statistical significance.

Discussion

The aim of this study was to determine the effect of a low-GI diet on PPH and postprandial glucose and insulin in individuals with chronic SCI. A lowGI breakfast tended to result in a lower drop in SBP in comparison to a high-GI meal under laboratory-controlled conditions. There was no difference in postprandial glucose or insulin iAUC, although the peak glucose and insulin concentrations were lower for the low-GI breakfast at 30-minutes post-meal consumption. In the home setting, there were no differences in the magnitude of decrease in SBP after eating between the two diets. However, peak postprandial glucose concentrations were lower at lunch and dinner during the low-GI compared to the high-GI diet.

The primary outcome measure was the maximum drop in SBP following breakfast under laboratory-controlled conditions. Seven of the 11 participants responded in the hypothesized direction, with the maximum drop in SBP lower in the low-GI compared to the high-GI diet. BP instability is common in individuals with SCI, particularly individuals with cervical level injuries,²⁸ and this may explain the variance observed in the present study. Despite not reaching statistical significance (P = .056), the medium effect size (d = 0.52) suggests there is a potential beneficial effect of a low-GI breakfast on the magnitude of PPH. On average, the low-GI breakfast resulted in a smaller drop in SBP compared to the high-GI breakfast by 10 mm Hg. Furthermore, the low-GI diet prevented PPH in one individual and substantially reduced the magnitude of PPH (24 and 50 mm Hg) in the other two individuals who had PPH during the experimental conditions. This may be clinically significant due to the dose-response relationship between the postprandial fall in SBP and mortality rate reported in older adults (>65 years) without SCI.¹⁰ These findings suggest that consuming low-GI meals may be recommended as a solution to reduce the magnitude of PPH in individuals with chronic SCI. As a caution to this, individuals rested in the supine position during this portion of the study, which has limited real-world applicability.

This is the first study to assess the efficacy of manipulating the GI of meals on PPH. Previous research in this area has focused on pharmacologic agents to treat PPH. For example, acarbose, an α-glucosidase inhibitor that decreases glucose absorption in the small intestine, has been shown to prevent PPH in two case reports in individuals with SCI.^29,30 Further, in a randomized clinical trial, in elderly individuals with autonomic dysfunction, acarbose administration (100 mg) reduced the magnitude of PPH (~17 mm Hg), with 10 mg/mg/dL lower glucose absorption and 11 μU/mL lower insulin secretion.²¹ Importantly, the reduction in PPH remained significant after the adjustment for insulin concentrations, suggesting additional mechanisms are also involved. In the present study, a reduction in the magnitude of PPH was achieved (10 mm Hg), with no difference in total insulin concentrations but a ~37% lower concentration at 30 minutes following the low-GI breakfast. Although there were no correlations between insulin concentrations and the magnitude of PPH, the study was not designed to determine whether the decrease in the maximum drop in SBP was caused by a reduction in insulin secretion. Importantly, acarbose has not been approved by the US Food and Drug Administration for treatment of PPH, and the interactions with other pharmacologic agents that individuals with SCI take are currently unclear. Therefore, this possible lifestyle solution is a promising solution to treat PPH.

In contrast to the findings under laboratory-controlled settings, there was no effect of diet on the magnitude of SBP drop after eating during the at-home phase of the study. There were difficulties in obtaining ambulatory BP measurements, resulting in 42.4% of meals being excluded from the analysis due to insufficient or missing data. This study applied strict criteria for analysis, with each meal requiring eight successful postprandial measurements to be included. The missing data are likely due to movement artefacts, despite clear instructions given to participants. Ambulatory BP monitoring in individuals with SCI provides additional challenges compared to individuals without SCI, including loss of readings due to increased reliance on upper extremities for wheelchair mobility and transferring.³¹ Furthermore, the sample largely consisted of individuals with tetraplegia, who may have been unable to adjust the cuff if it were temporarily out of alignment with the brachial artery. Research-grade wrist-worn devices capable of accurately measuring BP may be required to determine the efficacy of interventions on PPH in the at-home setting for this population. Despite this, 34.5% of successful recorded meals exhibited PPH. This prevalence is similar to previous research¹² and highlights the need for more research and preventative strategies to be developed for the condition.

There were no differences in postprandial glucose iAUC detected between diets in the inlab or at-home settings. However, peak glucose levels were lower at lunch (18 mg/dL) and dinner (23 mg/dL) with the low-GI compared to high-GI diet during the at-home setting. This effect was not present at breakfast, and the magnitude of the reduction in peak glucose appears to increase throughout the day; this may indicate a subsequent meal effect, whereby a previous low-GI meal (e.g., whole grains and legumes) slows absorption and digestion of starch to improve glucose tolerance in the following meal.³² This reduction in peak glucose levels may be clinically important as postprandial glucose peaks are a risk factor for CVD, even in individuals without diabetes.³³ Despite excluding individuals with diabetes treated with insulin, our sample included eight individuals with insulin resistance, two with prediabetes, and one with undiagnosed diabetes. Therefore, our findings are relevant to a wider population with increased risk of glycemic disorders.³⁴ Further, they support the need for studies examining the medium- to long-term effects of a low-GI diet on glycemic parameters in individuals with SCI, given evidence from non-SCI populations that low-GI diets are effective at reducing fasting glucose concentrations and glycated hemoglobin (HbA1c) in those with prediabetes and diabetes.³⁵

Limitations and adverse events

Firstly, due to participant drop-out, the target sample size of 12 was not reached; however, based on the observed effect size, an additional participant would not have meaningfully changed the main findings (maximum decrease in SBP during in-lab breakfast). Secondly, compliance (i.e., consuming all food prescribed) was lower than expected (81% for low GI and 90% for high GI). Two participants reported gastrointestinal side effects including nausea, diarrhea, and vomiting during the low-GI at-home phase, and they consumed a low percentage of their prescribed meals because of this (47% and 71%). This is likely due to the high fiber content of the low-GI meals, and future studies should aim to reduce this to increase real-world relevance. Additionally, only one participant fully consumed all the meals provided during the at-home phase. This suggests the calculated energy requirements of the sample may have been overestimated, and therefore future studies in this area should carefully consider the appropriate monitoring of physical activity energy expenditure, including wearable monitors during the measurement period.³⁶ Even though adjustment for compliance was performed, it is likely that the side effects and noncompliance to the low-GI diet affected glucose responses. Thirdly, individuals with and without PPH were included to maximize recruitment for this pilot study. The pre-experimental procedures included a PPH test to a standardized drink, and seven individuals met the criteria for PPH during this visit. Previous research has highlighted that one in every four meals met the threshold for PPH in individuals with SCI,¹² and therefore individuals were not excluded based on a one-off assessment. Finally, hydration status and physical activity were not recorded during either phase of the study, and this may have confounded SBP data. Water intake can invoke a sympathetic pressor response in individuals with SCI³⁷; drinking water before eating has been shown to attenuate PPH in individuals with autonomic failure.³⁸

Conclusion

The results from this randomized crossover study provide preliminary evidence that consuming low-GI meals may be an effective strategy to reduce the magnitude of PPH in individuals with chronic SCI, at least under laboratory-controlled conditions. The low-GI diet also reduced peak insulin and glucose concentrations, suggesting clinical indices of glycemic control may be improved with long-term adherence. These findings support the need for low-GI interventional studies with a larger sample size in individuals with SCI and PPH.

Funding Statement

J.S. received funding support from the National Institutes of Neurological Disorders-NIH (grant R01NS118200), the Craig H. Neilsen Foundation (CHNF#596764), the Wings-for-Life Spinal Cord Research Foundation (#DE-16/16), the Era-Net-NEURON Program of the European Union (EU) (SILENCE #01EW170A and SCI-Net #01EW1710), and the W.E. Hunt & C.M. Curtis Endowment. J.M.S. is a Discovery Theme Initiative Scholar (Chronic Brain Injury) of the Ohio State University.