Here's a little comparison of how UMTS and LTE deal with limited uplink power of mobile devices which I think it is quite interesting:
When uplink power for a UMTS E-DCH (HSUPA) transmission reaches a maximum, the number of simultaneously used codes can be reduced, a more conservative coding can be employed for additional redundancy and the modulation order can also be changed.
In LTE, modulation and coding can also changed as needed. And in addition, there's a third parameter: LTE uses an OFDM air interface, or to be more precise, SC-FDMA in the uplink direction. In other words, many subcarriers are used for the data transmission which are grouped into consecutive Resource Blocks (RBs) in case of uplink transmissions. When the mobile device reaches its maximum power level and the network detects this, it can reduce the number of RBs assigned in the uplink direction. This way the mobile can concentrate it's power on fewer RBs and hence it has more power available on the narrower channel it now uses. From a network point of view this is much better than leaving the number of RBs as they are and reduce modulation and coding as the RBs that are removed can be assigned to other devices also requesting resources in the uplink direction. For details see the power control section in this excellent book.