Many authors have reported on the successful use of transcutaneous oxygen measurements to determine the appropriate lower extremity amputation level [6–16]. One of the initial reports on this topic was by Franzeck et al. [7]. Mean tcpO₂levels in patients who experienced primary healing of a lower extremity amputation were compared to those of patients who failed to heal their amputation. The respective values for healing and nonhealing were 36.5  ±  17.5 and less than 30 mmHg. However, three of nine patients whose tcpO₂level was less than 10 mmHg healed primarily.

In a study of below-knee amputations, Burgess et al. [6] noted that all 15 amputations that were associated with a tcpO₂level greater than 40 mmHg healed. Primary wound healing was noted in 17 of 19 below-knee amputations with a tcpO₂measurement between 1 and 40 mmHg, but none of the three patients with a below-knee level of 0 mmHg healed their amputation. Katsamouris et al. [9] reported that lower extremity amputations healed in all 17 patients with a tcpO₂level greater than 38 mmHg or a pO₂index (chest wall control site) greater than 0.59. Ratliff et al. [11] reported that below-knee amputations healed in 18 patients with a tcpO₂measurement greater than 35 mmHg, whereas healing failed in 10 of 15 with a tcpO₂value less than 35 mmHg. In a study of 42 lower extremity amputations (28 below-knee and 14 above-knee), Christiansen and Klarke [14] found that 27 of 31 patients with a tcpO₂level greater than 30 mmHg healed primarily. Seven patients with values between 20 and 30 mmHg healed although four patients had delayed healing. The amputation stumps of all four patients with a value below 20 mmHg failed to heal because of skin necrosis.

Data from Wyss et al. [13] are comparable to those yielded by a prospective study evaluating multiple tests used for amputation level selection. In this study, tcpO₂measurements were prospectively compared to transcutaneous carbon dioxide tension, transcutaneous oxygen-to-carbon dioxide tension, foot-to-chest transcutaneous oxygen tension, intradermal xenon-133 clearance level, ankle brachial index, and the absolute popliteal artery pressure for accuracy in amputation level selection. All metabolic variables exhibited a high degree of statistical accuracy in predicting amputation healing, but none of the other tests showed statistical reliability. All amputations in this study (transmetatarsal, below-knee, and above-knee) healed primarily when the tcpO₂measurement was greater than 20 mmHg and there were no false-positive or false-negative results [10]. It was also noted that successful prediction of amputation healing for any of the metabolic parameters was not affected by the presence of diabetes mellitus. This finding is similar to the observation of Wyss et al. [13].

In contrast to lower extremity amputations in nondiabetics, which usually result peripheral vascular disease primarily, most amputations in diabetic patients result from various combinations of contributing causes including neuropathy, ischemia, alterations of white cell function, infection or gangrene, faulty wound healing, cutaneous ulceration, and minor trauma [15]. Malone et al. [10] and Christensen and Klarke [14] concluded that a transcutaneous oxygen tension of 20 mmHg or more accurately predicted amputation site healing and found no difference in the healing rate between diabetics and nondiabetics. Computerized analysis of various transcutaneous metabolic parameters by Malone et al. [10] demonstrated a high association with primary amputation site healing with the following values: transcutaneous oxygen tension greater than 20 mmHg, transcutaneous carbon dioxide value less than 40.5 mmHg, transcutaneous oxygen-to-transcutaneous carbon dioxide index greater than 0.472 and foot-to-chest transcutaneous oxygen index greater than 0.442.

The above data reinforce the fact that elective lower extremity amputation should not be performed without objective testing to ensure selection of the most distal amputation site that will heal primarily yet allow removal of infected, painful, or ischemic tissue [15, 16]. A variety of techniques are available to achieve this, depending on available equipment, the amputation level under consideration, and the accuracy of the chosen modality [16]. TcpO₂measurements continue to be a reliable technique; however, it is not suitable for whole limb mapping.

The ultimate role of any method used for amputation level determination is to inform the surgeon of the quantitative risk of nonhealing at the proposed site of surgery. The level of amputation can then be decided on the basis of this objective finding in conjunction with surgeon’s clinical judgment and patient’s physical findings. For example, a surgeon might perform an amputation distal to the knee through a site with a very low tcpO₂level in a patient who is well motivated, relatively young, and otherwise healthy. Such an amputation would almost certainly be ruled out in a fragile elderly person who faces a limited prospect for successful rehabilitation.

Successful treatment of the patient with diabetes and limb-threatening ischemia requires an accurate assessment of limb perfusion. Presenting clinical symptoms may be misleading. Physical examination of pedal pulses or ankle/brachial index (ABI) may not be accurate due to the non-compressible nature of a diabetic patient’s peripheral arteries. Often, the cause of the presenting foot problem is multifactorial and commonly used non-invasive lower extremity hemodynamic studies lack discriminative accuracy. On the other hand, arteriography is ultimately accurate. However, it is invasive, expensive, and carries a small but well-defined set of associated complications [17, 18]. In this setting, tcpO₂measurements can be extremely useful as they are noninvasive, inexpensive, and reproducible [19–24].

In a clinical experience reported by Ballard et al. [25], tcpO₂measurements were prospectively demonstrated to accurately predict severity of foot ischemia in patients with diabetes. Based on clinical experience and previously published amputation level determination data, an absolute transmetatarsal tcpO₂measurement of 30 mmHg was used as the threshold value for selection of a treatment option. If the level was 30 mmHg or greater, the patient’s foot problem was managed conservatively with local wound care, wound debridement, or minor foot amputation. If the level was less than 30 mmHg, arteriography of the involved limb was performed to plan arterial reconstruction or to perform percutaneous intervention to improve foot perfusion.

Thirty-one of 36 (86%) limbs in the conservatively managed group were treated successfully including 73% (11/15 ft) of limbs without a palpable pedal pulse. The mean time to wound healing was 6.85 weeks and there were five treatment failures. In the operative/endovascular group, 83.3% of limbs achieved a TM tcpO₂level ≥30 mmHg after treatment. Twenty-two of 26 (85%) limbs in this group had complete resolution of their presenting foot problem. The mean time to wound healing was 9.52 weeks. Treatment failures eventually led to 3 BKAs (1 failed necessitating revision to the AK level) and 1 above-knee amputation.

The pre-treatment pedal pulse examination was more accurate than an ABI in predicting forefoot tcpO₂values above or below 30 mmHg. Further, an abnormal arteriogram was predicted by both a low TM tcpO₂level and the absence of a palpable pedal pulse, but not by an ABI <60. The presence of a pedal pulse was 100% accurate for identifying limbs with a TM tcpO₂ ≥  30 mmHg, but there were an additional 17 limbs with a measurement ≥30 mmHg and no palpable pedal pulse. Following arterial bypass or angioplasty, a TM tcpO₂level ≥30 mmHg was highly accurate in predicting a successful outcome. Ultimately, an initial or post-intervention TM tcpO₂level ≥30 mmHg was more accurate than a palpable pulse in predicting either wound healing or resolution of rest pain. An ABI  ≥  0.60 was also associated with a successful outcome, but due to non-compressible vessels, this was only able to be calculated in 41/62 (66%) limbs.

Certainly diabetic patients without pedal pulses do have arteriosclerotic lesions, some of which can be reconstructed. However, this prospective study demonstrated that such surgical revascularization is not obligatory. In fact, well-performed tcpO₂measurements predicted distal ischemic wound healing in 90% of cases. Furthermore, conservative management was not only cost effective when compared to surgical or endovascular revascularization, but time to wound healing was not statistically significantly different between the two groups (6.84 weeks versus 9.52 weeks, P =  0.169).

As demonstrated in the study outlined above, an absolute TM tcpO₂level ≥30 mmHg appears to be an accurate cutoff point for the selection of treatment for almost all diabetic foot problems. The conservative management scheme, however, requires diligent patient follow-up. There must be a commitment by the surgeon to perform wound debridements and staged procedures (i.e., minor foot amputations or split-thickness skin grafts). Proper outpatient wound care is essential. Finally, a higher TM tcpO₂threshold (40 mmHg) should be used to select management of calcaneal gangrene or extensive non-healing ulcerations. Table 52.1demonstrates an algorithm for the elective management of diabetic patients with limb-threatening ischemia based on tcpO₂level.