To give a short answer,(and sorry I'm not an expert!) I don't think we actually know! Though most people would suggest that the fewer high blood sugar excursions the better.
The biggest short term danger from high blood glucose levels is DKA. At 15mmol you should perhaps be checking for ketones and correcting to bring the levels and any ketones down.
However 10 until very recently was the top bracket of postprandial targets, most authorities now give a lower figure.The UK after meal type 1 target is now a maximum of 9 mmol. I
I'm not certain there has been any long term research into the effects of day to day variations or odd incidences of hyperglycaemia. It is all pooled together to show the effects of overall control.
Much of our data on complications comes from the DCCT
The DCCT trials showed that elevated HbA1cs over time are directly correlated with complications. The higher the average HbA1c the sooner the likelihood of complications (particularly microvascular).
http://www.dgdiabetes.scot.nhs.uk/icp/icptype2/hba1ctargets.shtml
A 2008 reanalysis of the data suggests that variable HbA1cs, that is short-term glucose instability can have an adverse effect.
http://www.ncbi.nlm.nih.gov/pubmed/18650371?dopt=AbstractPlus
Obviously to have a consistently lower HbA1c it is necessary to have lower overall levels but the odd higher level, brought down quickly wouldn't be reflected in the HbA1c. (My HbA1c is consistently in the 5s and I rarely go above 8mmol but it does sometimes happen)
recent studies have found high er post prandial levels to be predicitve of CV events in type
2 but I've found nothiing directly on type 1.
http://diabetes.diabetesjournals.org/content/54/1/1.abstract
http://jcem.endojournals.org/cgi/content/abstract/91/3/813