Significant Time Until Catheter Occlusion Alerts in Currently Marketed Insulin Pumps at Two Basal Rates

Abstract

Dear Editor:

O ne of the most common problems of continuous subcutaneous insulin infusion (CSII) is occlusion of the infusion set. Frequencies up to 30% have been reported in literature, although objective assessment of occlusion is difficult.^1,2 CSII devices feature an occlusion detection alert based on a sensor connected to an electronics system that measures pressure, compression, or displacement of the lead screw in the pump drive system. Above a certain threshold, an alert is given. A relatively large displacement of the lead screw may be needed to activate the alert; in other words, it may take quite some time before the occlusion alert is activated. Some CSII manufacturers provide information in the manual on the time needed to activate the alert after total occlusion. For example, the time until occlusion alert in the D-Tron™ plus (Disetronic Medical Systems, Member of the Roche Group, Basel, Switzerland) is given as between 2 and 4 h at a basal rate of 1.0 IU/h. If the basal rate decreases, the occlusion alert will be activated later. As an illustration, the time until occlusion alert of the Accu-Chek^® Spirit Combo (Disetronic Medical Systems) is given as shorter than 5 h at a basal rate of 1.0 IU/h, whereas at a basal rate of 0.1 IU/h, the alert will go off after up to 50 h. An occlusion alert is especially important during the night; therefore we examined the time until occlusion alert at basal rates.

Five commercially available devices were examined: Paradigm^® Veo™ (Medtronic Minimed, Minneapolis, MN), D-Tron, Spirit Combo, Dana (Sooil, Seoul, Korea), and OmniPod^® (Insulet Corp., Bedford, MA). The catheter was occluded completely applying a clip at the tip of the catheter at t=0. Two commonly used CSII delivery rates were chosen: 1.0 IU/h and 0.5 IU/h. Insulin aspart was used in all pumps. The fluid pressure in the catheter depends on its compliance and thereby on its length, and therefore two different lengths of the catheter were tested: 60 cm and 2.5 cm. The OmniPod was only tested with a length of 2.5 cm. All tests were performed in duplicate except for the OmniPod, for which testing was done three times at a basal rate of 1.0 IU/h and three times at 0.5 IU/h. It was assured that no insulin came out of the catheter during the observation time.

The results are given in Table 1. Time to occlusion alert varied from 1 h 52 min to 4 h 32 min at a basal rate of 1.0 IU/h with infusion catheters of 60 cm. At a basal rate of 0.5 IU/h the occlusion alert time varied from 1 h 34 min to 9 h 25 min. No large differences were seen between basal rates in the Paradigm Veo and Spirit Combo. If the length of the infusion set was 60 cm with a basal rate of 0.5 IU/h, the time to occlusion alert doubled in Dana and almost tripled in D-Tron compared with a basal rate of 1.0 IU/h. The alert in the Dana pump has three sensitivity levels: low, medium, and high. Increasing the sensitivity from medium to high decreased the time until an alert was given. If the infusion set was occluded at 2.5 cm with a basal rate of 1.0 IU/h, the occlusion alert occurred faster in three of four insulin pumps compared with 60 cm. If the insulin rate was 0.5 IU/h, two devices had an earlier alert with an occlusion of 2.5 cm compared with the infusion set of 60 cm. Overall, the OmniPod seemed to need more time to activate the occlusion alert than the other pumps.

Table 1.

Time to Occlusion Alert in Different Commercial Devices with a Length of the Infusion Set of 60 or 2.5 cm

		Basal rate, catheter length
Device (manufacturer), stepsize	Manufacturer alert	1.0 IU/h 60 cm	0.5 IU/h 60 cm	1.0 IU/h 2.5 cm	0.5 IU/h 2.5 cm
Paradigm Veo (Medtronic Minimed), step size 0.05	Basal rate 1.0 IU/h: 2 h 25 min–3 h 9 min	1 h 55 min 1 h 56 min	1 h 59 min 1 h 34 min	44 min 27 min	1 h 47 min 1 h 38 min
	Basal rate 0.025 IU/h:
	66–189 h
D-Tron plus (Roche Diagnostics), step size 0.1	Basal rate 1.0 IU/h:	3 h 17 min	9 h 25 min	3 h 50 min	4 h 22 min
	2–4 h	2 h 45 min	8 h 46 min	3 h 6 min	4 h
Accu-Chek Spirit Combo (Roche Diagnostics), step size 0.05	Basal rate 1.0 IU/h: ≤5 h	1 h 51 min	2 hr 7 min	56 min	1 h 54 min
	Basal rate 0.1 IU/h: ≤50 h	2 h 6 min	1 h 58 min	57 min	1 h 58 min
Dana Diabecare R (Sooil), step size 0.1	No information about occlusion time	Sensitivity:	Sensitivity:	Sensitivity:	Sensitivity:
		Medium:	Medium:	Medium:	Medium:
		4 h 32 min	8 h 21 min	2 h 9 min	6 h 16 min
		2 h 20 min	5 h 15 min	1 h 54 min	5 h 35 min
		High:	High:	High:	High:
		2 h 5 min	4 h min	55 min	1 h 25 min
		1 h 53 min	3 h 29 min	34 min	52 min
OmniPod (Insulet Corp.), step size 0.05	No information about occlusion time			7 h 20 min	>24 h
				9 h 17 min	>24 h
				5 h 29 min	>24 h

In conclusion, the time until occlusion alert given in the manufacturers' manuals was somewhat longer than we found in our study. In three of the four devices the time until occlusion alert at a catheter length of 2.5 cm was shorter than with a catheter length of 60 cm, if the basal rate was 1.0 IU/h. This is in accordance with the fact that the distention capacity of the 2.5-cm catheter is less than that of the 60-cm catheter. However, it remains unclear why this was not seen in the fourth pump. At a basal rate of 0.5 IU/h, the time until occlusion alert at a catheter length of 2.5 cm was shorter in two of the four devices compared with 60 cm. An explanation for our finding that the alert did not occur faster in the other two devices despite the shorter catheter length could be that the displacements of the lead screw are too small to build a threshold pressure to trigger the alert at this low basal rate. The dislocations of the screw lead are equivalent to the step sizes of insulin infusion. In the two devices with an earlier alert, the step size is bigger than in the other two devices: 0.1 compared with 0.05 IU of insulin. Therefore, in these devices, the threshold pressure is built up earlier and the occlusion is detected faster.

Thus the time until an occlusion alert differed per basal rate, length of the infusion catheter, and the commercial device, although the pattern is not always completely predictable or consistent. Efforts from manufacturers to decrease the time between occlusion and alarm would be welcome. For the time being patients and caregivers should realize that occlusion alerts are not very sensitive.

References

Raskin

, Holcombe

, Tamborlane

, Malone

, Strowig

, Ahern

, Lavent

. A comparison of insulin lispro and buffered regular human insulin administered via continuous subcutaneous insulin infusion pump. J Diabetes Complications, 2001; 15:295–300.

van Bon

, Bode

, Langeron

, DeVries

, Charpentier

. Insulin glulisine compared to insulin aspart and to insulin lispro administered by continuous subcutaneous insulin infusion in patients with type 1 diabetes: a randomized controlled trial. Diabetes Technol Ther, 2011; 13:607–614.