The recent review by Koelen and colleagues ^{Reference Koelen, Houtveen, Abbass, Luyten, Eurelings-Bontekoe and Van Broeckhuysen-Kloth1} of psychotherapy for severe somatoform disorder is welcome in highlighting the need for better evidence in this area. It has unfortunately omitted a number of relevant studies, especially relating to conversion disorder. One major reason for this is that the index date on which studies were searched for, March 2010, was nearly 4 years prior to publication. It is a pity that the authors did not update their analysis at the time of their last revision in June 2013, as they would, at this time, have been able to include a number of relevant studies, including a randomised trial of cognitive-behavioural therapy for non-epileptic seizures (n = 66) ^{Reference Goldstein, Chalder, Chigwedere, Khondoker, Moriarty and Toone2} and a randomised controlled trial of guided self-help for functional neurological symptoms (i.e. conversion disorder) (n = 127). ^{Reference Sharpe, Walker, Williams, Stone, Cavanagh and Murray3} These two studies were published before one of the studies included in the analysis, the study by Sattel et al published in 2012. ^{Reference Sattel, Lahmann, Gündel, Guthrie, Kruse and Noll-Hussong4}

There are further studies of psychotherapy in conversion disorder which were published before March 2010: a study of psychotherapy for non-epileptic seizures (n = 20); ^{Reference Goldstein, Deale, Mitchell-O'Malley, Toone and Mellers5} a study of psychotherapy for conversion disorder (n = 91); ^{Reference Reuber, Burness, Howlett, Brazier and Grunewald6} a study of psychotherapy for psychogenic movement disorders (n = 10); ^{Reference Hinson, Weinstein, Bernard, Leurgans and Goetz7} and a large controlled and negative trial of psychotherapy for patients with somatoform disorders in a general hospital (n = 91). ^{Reference Schweickhardt, Larisch, Wirsching and Fritzsche8} The authors may have excluded them but they did not present a list of the 64 excluded studies as a supplemental file.

Other types of study that could arguably have been included using the authors’ own criteria are some randomised trials in functional dysphonia, a form of conversion disorder treated in secondary care with voice therapy and sometimes psychotherapy. ^{Reference Ruotsalainen, Sellman, Lehto, Jauhiainen and Verbeek9} There are also treatment studies of children with conversion disorder which have not been included and would not have been excluded by the authors’ inclusion criteria. ^{Reference Turgay10,Reference Schwingenschuh, Pont-Sunyer, Surtees, Edwards and Bhatia11}

Further studies in conversion disorder have followed in the past 2 years which describe outcomes from multidisciplinary treatment including psychotherapy. ^{Reference Aybek, Hubschmid, Mossinger, Berney and Vingerhoets12-Reference Kompoliti, Wilson, Stebbins, Bernard and Hinson15} Journal articles cannot always be up to date, but the number of omissions here make this meta-analysis immediately in need of updating.

Two included studies were of hypnosis for motor conversion disorder. ^{Reference Moene, Spinhoven, Hoogduin and Van Dyck16,Reference Moene, Spinhoven, Hoogduin and Van Dyck17} Hypnosis is arguably a form of psychotherapy, but also arguably not. In addition, the inclusion of studies which randomised bioenergetic exercise against gym exercise in a setting where all patients received psychotherapy ^{Reference Nickel, Cangoez, Bachler, Muehlbacher, Lojewski and Mueller-Rabe18} and a study of in-patient multidisciplinary rehabilitation in chronic pain (n = 298) graded as ‘extremely poor’ ^{Reference Schneider and Rief19} and then included in a ‘treatment as usual arm’ is debatable.

The authors could have done more to highlight one of the obvious drawbacks of their review. There is a paradox in reporting on treatment for patients who had been defined as having somatoform disorder (often needing only to have three symptoms e.g. pain, fatigue, dizziness or irritable bowel syndrome) while ignoring studies on psychotherapy for individual functional somatic disorders such as irritable bowel syndrome and fibromyalgia. Most patients with functional somatic disorder also have other symptoms such as fatigue and pain, ^{Reference Wessely, Nimnuan and Sharpe20} and probably would, for example, meet criteria for multisomatoform disorder. It is at times highly arbritary whether authors decide, for example, to use the term somatoform pain disorder or chronic pain disorder. A broader overview of studies in all these fields or at least greater acknowledgement of the overlap would have been helpful for the reader.