No difference in pain relief between 1kHz and 10kHz spinal cord stimulation in PROCO RCT


A randomised controlled trial has found excellent pain relief and no clinical difference among spinal cord stimulation frequencies from 1kHz–10kHz. Further, the study showed that 1kHz stimulation provides similar pain relief using significantly less energy than higher frequencies.

The PROCO (Effects of Pulse Rate On Clinical Outcomes in Kilohertz Frequency Spinal Cord Stimulation) randomised controlled trial is a multicentre, double-blind, crossover study. The results were presented at the International Neuromodulation Society World Congress (INS; 27 May–1 June, Edinburgh, UK).

According to the study’s chief investigator, Simon Thomson (consultant in Pain Medicine and Neuromodulation at Basildon and Thurrock University Hospitals NHS Foundations Trust, Basildon, UK), the rationale for the trial came about when excellent results were observed in studies investigation 10kHz spinal cord stimulation. “We all asked: Why 10kHz? Is there a clinical difference in pain relief using spinal cord stimulation frequencies from 1kHz to 10kHz? Previous studies showed that SCS can be effective at 1kHz and 10kHz,” he explained.

The study design was based on usual care at Basildon hospital—lead placement spanning T9 to T10 vertebral region, verified paraesthesia coverage ≥80% of pain areas— each participant had a baseline pain recording period before having a paraesthesia based spinal cord stimulation trial. If the patient achieved 50% or more pain reduction as judged by the hospital’s standard of care, a permanent implant was performed and stimulation was turned off.

As part of the trial, the patients used a validated electronic diary (e-diary). The e-diary was in the form of a watch strap worn by the patients for nine months that would prompt them three times a day to enter a pain score from three sites—the back, legs and overall pain; it also confirmed that their opioid dose had remained stable.

“We used an electronic diary because we are aware that the paper based diary system can have as little compliance as 11%, and when the patients come back for their data collection, they are often asked in retrospect for their pain score over the past five days. The e-diary prompted each patient for 180 pain scores over the randomisation phase—three pain areas (back, leg, overall), three times a day, five days per rate evaluation period and four rates. The e-diary pain scores averaged three per day over five days translating to 15 data points per patient per rate evaluation period,” Thomson explained.

Patients included in the study had persistent lower back pain (≥ leg pain) for >90 days, with NRS ≥5 for lower back pain. The patients were also stable on opioid medications for 30 days prior to screening and had a baseline Oswestry Disability Index score of ≥20 and ≤80. Patients were excluded if they had had back surgery in the previous six months, if they had a confounding medical or psychological conditions or untreated major psychiatric comorbidity or serious drug-related behaviour issues.

A key component of the study was the “sweet spot” search that was done at 10kHz. The bipoles varied over that up to an eight-week period between T8 and T11. Investigators always started at the T9–10 interspace and moved accordingly until the sweet spot was identified based on the best pain relief. For scientific purposes, the study afforded the opportunity to exhaustively test stimulation locations along both leads—up to 14 different stimulation locations could be tested in up to eight weeks and the amplitude was optimised at each stimulation location for each patient. Only those patients who were 10kHz responders (minimum of >30% pain reduction as per e-diary) continued in the study.

In the randomisation phase, each patient experienced each rate (1kHz, 4kHz, 7kHz and 10kHz) for four weeks. The order of delivery of the different rates was ascertained by a random computer generator. During each period, the pulse width and the amplitude were optimised at each frequency for each patient and there was a wash-out period between each rate randomisation phase. The wash-out period was defined as complete when their pain scores had risen to within 80% of their baseline.

In the study, 20 patients completed randomisation and have follow-up data. The patients were 11 male, 32–75 years old (mean=53 years), 16 with failed back surgery syndrome, four with chronic radiculopathy, and a pain duration of one to 27 years (mean=11 years).

Simon Thomson

Thomson reported that in terms of back pain, the mean pain score as judged by the e-diary was approximately 7/10. At 10kHz the investigators observed an excellent response with a greater than 50% reduction in pain. At 7kHz the results were about the same. At 4kHz the reduction in pain also remained about the same. And finally, 1kHz stimulation provided also equivalent pain relief to all of the higher frequencies (p=0.00002).

As it relates to leg pain, all four frequencies also provided equivalent improvement in pain scores (p=0.003). Overall pain results were also similar, with excellent pain relief across the randomisation spectrum (p=0.00002).

When it came to secondary endpoints of quality-of-life measurements, there was again no statistical difference between frequencies (p>0.8).

The main difference observed among the different frequencies was in charge per second where 1kHz stimulation was around three times more efficient than 10kHz both at optimal neural dose (p=0.000002).

Thomson explained that neural dosing is key and its goal is to achieve optimal clinical outcome at minimal dose. For this, the pulse width and amplitude must be optimised for a given frequency and it is not as simple as holding charge per time constant.

“Achieving pain relief requires delivering the right waveform to the right target, very careful finding of this bipole to achieve best pain relief and then optimisation of the stimulation by amplitude and pulse width at each frequency. Frequency cannot be looked at in isolation—the different frequencies required different pulse width and amplitude combinations,” he said.

Finally, Thomson reported that the PROCO study demonstrates that there is no statistical difference between 1kHz and 10kHz stimulation using optimal neural dosing. This is consistent with preclinical studies from Shechter et al (2013) as well as Song et al (2014).

“In conclusion, the PROCO RCT showed no clinical difference in pain relief using frequencies from 1–10kHz, and we believe that proper neural dosing is required to optimise frequencies. Further, 1kHz provides excellent pain relief using significantly less energy,” he stated.

NeuroNews caught up with Simon Thomson to get more insight on the potential implications of the PROCO study and the questions that remain unanswered.

What implications could the results of the PROCO study have on current clinical practice?

There is no doubt that kHz frequency stimulation of the spinal cord can achieve pain relief. There is no superiority of 10kHz stimulation over 1khz stimulation provided that there is targeting and optimal neural dosing of the stimulation. Patients can enjoy the benefits of both paraesthesia based SCS and kHz frequency SCS from the same device.

What does the future hold in terms of further research exploring the rate of pain relief with different SCS frequencies?

The PROCO RCT is just scratching the surface of our understanding. This experiment just used a bipole and set out to find the 10kHz sweet spot. But is the 10kHz sweet spot the same as a 1kHz sweet spot? Would a more complex field shape with multiple cathodes and anodes whilst using kHz SCS be more effective? How low can the frequency go at sub-perception stimulation?

The study found that 1kHz stimulation provided similar pain relief that higher frequencies using significantly less energy. Why is this important?

kHz frequency stimulation requires high-energy consumption. As a result rechargeable devices are used but may need to be recharged daily. The minimal electrical neural dose, MEND, is the minimum charge per second required to achieve at least 50% reduction in reported pain when compared to baseline. This figure will vary according to which kHz frequency and target array. Using these principles will help patients reduce the burden of device recharging.

If lower frequencies provide similar pain relief, is there a need to use higher frequencies at all?

The answer to this question will depend upon the ease of optimisation of the neural dose at each kHz frequency. We had some knowledge when we started this project, we learnt a lot, but there is so much more to learn. We hope that automation of the targeting and neural dose optimisation will make this process so much easier and efficient.


Please enter your comment!
Please enter your name here