Environmental effects of ambient temperature and relative humidity on insulin pharmacodynamics in adults with type 1 diabetes mellitus

This study aimed to explore the effects of ambient temperature and relative humidity on insulin pharmacodynamics in adults with type 1 diabetes.

Objective: This study aimed to explore the effects of ambient temperature and relative humidity on insulin pharmacodynamics in adults with type 1 diabetes.
Results: Higher temperature (30 C) under 10% fixed humidity conditions resulted in greater GIR max (P = 0.04) and a later t GIR.max (P = 0.049) compared to lower temperature (15 C). Humidity did not affect any pharmacodynamic parameter. When the combined effects of temperature and humidity were explored, t GIR.max (P = 0.008) occurred earlier, with a lower late insulin pharmacodynamic effect (AUC GIR.2-6h ; P = 0.017) at a temperature of 15 C and humidity of 10% compared to a temperature of 30 C and humidity of 60%.
Conclusions: High ambient temperature resulted in a greater insulin peak effect compared to low ambient temperature, with the contribution of high relative humidity apparent only at high ambient temperature. This suggests that patients with type 1 diabetes mellitus who are entering higher environmental temperatures, with or without high humidity, could experience more hypoglycaemic events. shown to accelerate insulin absorption and improve insulin sensitivity in patients with diabetes, with these effects largely mediated by an increase in skin temperature which results in an increased perfusion at the injection site.
The effects of relative humidity on insulin pharmacodynamics and pharmacokinetics are largely unexplored. An epidemiological study conducted in the Mediterranean area suggested an increased prevalence of diabetes among the elderly population on islands with high relative environmental humidity when adjusted for ambient temperature. 16 Notably, high relative humidity often occurs in the presence of high ambient temperature, making it challenging to unravel their individual effects. 17 Individuals with diabetes appear to tolerate moist, warm air with more than 50% humidity less well than adults without diabetes. 18 This may be explained by the fact that high humidity, when combined with high temperature, decreases the rate of cooling in the human body, leading to tiredness, exhaustion, reduction in alertness and, potentially, heat stroke, 17,19,20 which may also affect glycaemic control.
In order to assess the independent and combined effects of ambient temperature and relative humidity, this study evaluated the insulin pharmacodynamic profile following a single injection of a short-acting insulin analogue.

| Euglycaemic glucose clamp procedure
Prior to the euglycaemic glucose clamp procedure, all participants fasted overnight and for the duration of the six-hour procedure.
Water was allowed as required. In the clinic room, with the participant in a comfortable supine or semi-supine position, vital signs were recorded before two cannulas were inserted. One was inserted into the hand or forearm for venous sampling, with the hand heated to 55 C throughout the clamp procedure, allowing arterialization of the venous blood. 21 The second was inserted into the opposite arm, at the cubital fossa, for a variable infusion of insulin (15 units  In the environmental chamber, participants were instructed to wear light clothes to mimic real-life situations. The variable glucose infusion was used to maintain the target blood glucose level of 5.5 mmol/L (100 mg/dL) AE 20% guided by an algorithm 22 and by the participants' blood glucose concentration, measured within the preceding 5 minutes. Blood glucose concentrations were measured using a glucose analyser (HemoCue glucose 201+, Radiometer Ltd, Crawley, UK) and were recorded along with the glucose infusion rate every 5-10 minutes throughout the clamp procedure. Upon completion of the clamp procedure, vital signs were assessed and lunch was provided before discharge.

| Statistical analysis
The exogenous glucose infusion rate (GIR) was analysed every 5 to 10 minutes throughout the clamp procedure. A weighted local regression technique (LOESS) with a smoothing factor (SF) of 0.1 for calculation of time-related parameters and maximum GIR in accordance with previous studies that investigated the pharmacodynamics of short-acting insulin. 23 The pharmacodynamic endpoints calculated for each clamp study visit (Visits 2b, 3 and 4) were maximum glucose infusion rate (GIR max ) and time to maximum glucose infusion rate (t GIRmax ). In addition to total area under the curve (AUC) for GIR from 0-6 h (AUC GIR.0-6h ), partial AUCs from 0-1, 0-2 hours (AUC GIR.0-1h ), 0-6 hours (AUC GIR.0-2h ) and 2-6 hours (AUC GIR.2-6h ) following insulin injection were also calculated to determine early and late insulin action. A two-way ANOVA with temperature, humidity and their interaction as fixed effects and the participant as random effect was used for determination of AUC GIR.0-1h , AUC GIR.0-2h , AUC GIR.0-6h , AUC GIR.2-6h , GIR max (SF = 0.1) and t GIR.max(SF = 0.1) .
Data are presented as mean (1SD) and statistical significance was set at P ≤ 0.05. For graphical presentation (Figure 1) an SF of 0.3 was used and 10 data points with GIR-values of nearly 40 mg kg −1 min −1 in one participant were excluded in order to minimize random GIR-fluctuations.
Statistical analysis was conducted using SAS, version 9.4.

| RESULTS
Demographic and clinical characteristics of participating adults with type 1 diabetes mellitus at baseline are presented in Table 1.

| Independent effects of ambient temperature
As illustrated in Figure 1 and Table 2, at a temperature of 30 C with 10% humidity, the time-action curve of insulin was shifted to the right, with a later t GIR.max (P = 0.049) and a significantly greater GIR max (P = 0.04), compared to the condition at a temperature of 15 C and the same level of humidity, 10%. Although AUC GIR.0-1h, AUC GIR.0-2h and AUC GIR.0-6h did not differ significantly between the conditions with different temperatures, there was a trend towards higher AUC GIR 2-6h when comparing conditions at 30 and 15 C (P = 0.08) ( Table 2).

| Independent effects of relative humidity
There was no effect of humidity on insulin pharmacodynamics, as indicated by the absence of significant differences in GIR max , t GIR.max and AUCs for the time-action profile between the condition at 30 C with 10% humidity and the condition at 30 C with 60% humidity (P values between 0.21 and 0.95) ( Table 2).  Table 2) with less glucose to be infused at a lower temperature and humidity, but no differences were seen in early effects (AUC GIR.0-1h , P = 0.48 and AUC GIR.0-2h , P = 0.87) and overall effects (AUC GIR.0-6h , P = 0.48) on insulin action ( Table 2).

| DISCUSSION
Using the glucose clamp technique, the present study demonstrated that sudden changes in environmental conditions affect short-acting insulin analogue (insulin lispro) pharmacodynamics in adult men with type 1 diabetes mellitus. In response to higher temperature (30 vs 15 C) with fixed humidity, there was greater GIR max and a trend towards greater AUC GIR2-6h . High humidity affected insulin pharmacodynamics only when it was combined with high temperature. The mean time to GIR max was prolonged at 30 C with 10% or 60% humidity compared to 15 C with 10% humidity, and GIR max and late AUC (AUC GIR2-6h ) were greater, suggesting enhanced insulin absorption and peak effect. and those at 30 C concerning AUC GIR 0-6h , but there was a trend towards greater AUC GIR 0-6h with a higher temperature. The same study 7 also assessed insulin pharmacokinetic parameters and showed a three-to five-fold higher AUC for plasma-free insulin at 30 C than at 10 C, regardless of exercise. We cannot provide comparative data on these aspects, given that our study is limited to insulin pharmacodynamics and does not include a pharmacokinetic profile. Furthermore, it is more challenging to detect differences in pharmacodynamic parameters than in pharmacokinetic parameters, as the former are often characterized by greater variability and, therefore, pharmacokinetic results would be expected to be in line with pharmacodynamic findings in our study.   25 It is speculated that the marked effects of exposure to cold may be the result of enhanced insulin sensitivity and/or increased responsiveness for glucose uptake in peripheral tissues such as skeletal muscles. [25][26][27] However, in the current study we cannot provide further insight into these mechanisms, given that subcutaneous insulin was used and, therefore, other factors (eg, visceral and subcutaneous tissues) may have differentially affected the pharmacodynamic parameters.
Short term exposure to different levels of relative humidity, 10% and 60%, under fixed temperature had no effect on the insulin timeaction profile. However, exposure to high relative humidity in combination with high ambient temperature resulted in prolonged time to GIR max and a greater insulin pharmacodynamic effect compared to responses to the low temperature-low humidity condition, suggesting that high humidity may augment the high-temperature effect on enhanced insulin absorption from the injection site, but has little effect in its own right.
In conclusion, high ambient temperature resulted in greater insulin peak effect compared to low ambient temperature, with the contribution of high relative humidity to insulin absorption apparent only at high ambient temperature. This suggests that patients with type 1 diabetes mellitus who are entering an environment with higher temperatures, with or without high humidity, could experience more hypoglycaemic events.

ACKNOWLEDGMENTS
We thank all study participants for their commitment to this study.

CONFLICTS OF INTEREST
No potential conflict of interest relevant to this article was reported.